{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "63555cb17d7e92951d8919c8", "attempt_id": "02b0af4db4f7", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:51:59.341925Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 2342, "message_tokens": 2256, "message_utf8_bytes": 8198, "cache_tensor_bytes": null, "serialized_payload_bytes": 8198, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.00300002284348011, "ttft_ms": 14.371374971233308, "decode_ms": 120.50799990538508, "receiver_prefill_ms": 282.24879200570285}, "total_wall_ms": 425.71774998214096, "ttft_ms": 14.371374971233308, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "9,220,745", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "8af6edf6736154e1eb1c85c2253cd550117bde6765534baeea96c12d635d74d1", "output_hash": "ae55ba88b2cb1b0f1a4e0960ae312e0a1c177c79a6d14b67bff1677cec19b649"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "ca8813b5091c2635b61a70fe", "attempt_id": "365e40f0e039", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:51:59.784336Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 2344, "message_tokens": 2256, "message_utf8_bytes": 8198, "cache_tensor_bytes": null, "serialized_payload_bytes": 8198, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.00300002284348011, "ttft_ms": 13.773957965895534, "decode_ms": 134.89341700915247, "receiver_prefill_ms": 260.00020804349333}, "total_wall_ms": 412.1776249958202, "ttft_ms": 13.773957965895534, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "LY-51E-3199", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "8af6edf6736154e1eb1c85c2253cd550117bde6765534baeea96c12d635d74d1", "output_hash": "a1245d1536890a7e4b026366df48441d503f491f782cfd388797b5782443c5c7"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "af0c7358b8a7012a9d861be1", "attempt_id": "e3ce7935ac9c", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:52:00.210933Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 2348, "message_tokens": 2256, "message_utf8_bytes": 8198, "cache_tensor_bytes": null, "serialized_payload_bytes": 8198, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.00300002284348011, "ttft_ms": 13.239875086583197, "decode_ms": 93.6063330154866, "receiver_prefill_ms": 258.78491706680506}, "total_wall_ms": 369.173499988392, "ttft_ms": 13.239875086583197, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "2126290", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "8af6edf6736154e1eb1c85c2253cd550117bde6765534baeea96c12d635d74d1", "output_hash": "a030b162f5adef1d9151ead9ebe84b7d516e2701c0e6988ecf196d2652bfd94e"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a7a9fb3649b5234a9d3b6560", "attempt_id": "d30df1c16582", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:52:00.594948Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 2344, "message_tokens": 2256, "message_utf8_bytes": 8198, "cache_tensor_bytes": null, "serialized_payload_bytes": 8198, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.00300002284348011, "ttft_ms": 13.68141605053097, "decode_ms": 119.83075004536659, "receiver_prefill_ms": 258.1722919130698}, "total_wall_ms": 395.31645900569856, "ttft_ms": 13.68141605053097, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1,034,805", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "8af6edf6736154e1eb1c85c2253cd550117bde6765534baeea96c12d635d74d1", "output_hash": "b1dc6064f97c41a63e2b857664a4196c5397f8224728e25ada16fcf249b7ed60"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "4a4b8b9f897b8b379fa0ec7c", "attempt_id": "840cc60fe69d", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:52:01.004746Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 2350, "message_tokens": 2256, "message_utf8_bytes": 8198, "cache_tensor_bytes": null, "serialized_payload_bytes": 8198, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.00300002284348011, "ttft_ms": 13.237625011242926, "decode_ms": 66.63216603919864, "receiver_prefill_ms": 259.42050002049655}, "total_wall_ms": 342.75120799429715, "ttft_ms": 13.237625011242926, "input_usage": {}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Felaldnak-7", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "8af6edf6736154e1eb1c85c2253cd550117bde6765534baeea96c12d635d74d1", "output_hash": "2b9a4cf7221f1b6f31508a6b3078ec521f211026a932da5020cddd512ecc5fa2"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "bf2c8b49aa7a040c093f9638", "attempt_id": "d940fe66bec6", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:52:01.362565Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 2345, "message_tokens": 2256, "message_utf8_bytes": 8198, "cache_tensor_bytes": null, "serialized_payload_bytes": 8198, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.00300002284348011, "ttft_ms": 13.601499958895147, "decode_ms": 133.5405829595402, "receiver_prefill_ms": 259.28583403583616}, "total_wall_ms": 409.8432910395786, "ttft_ms": 13.601499958895147, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "14,910,510", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "8af6edf6736154e1eb1c85c2253cd550117bde6765534baeea96c12d635d74d1", "output_hash": "7e3d88ae4c83ccaf4ef5eff1e74cf595376f297b9fec58bbea6797c6bb5b0eb8"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "9b6f605520f967f369a7e683", "attempt_id": "1ad8ee01d2a3", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:52:01.787131Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 2345, "message_tokens": 2256, "message_utf8_bytes": 8198, "cache_tensor_bytes": null, "serialized_payload_bytes": 8198, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.00300002284348011, "ttft_ms": 13.508207979612052, "decode_ms": 13.383082929067314, "receiver_prefill_ms": 258.5025840671733}, "total_wall_ms": 288.84041705168784, "ttft_ms": 13.508207979612052, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "8af6edf6736154e1eb1c85c2253cd550117bde6765534baeea96c12d635d74d1", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "8eaff20b3af3c6b8fdbd0356", "attempt_id": "964f7c2436ba", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:52:02.090117Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 2347, "message_tokens": 2256, "message_utf8_bytes": 8198, "cache_tensor_bytes": null, "serialized_payload_bytes": 8198, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.00300002284348011, "ttft_ms": 13.426082907244563, "decode_ms": 13.3943340042606, "receiver_prefill_ms": 259.2853750102222}, "total_wall_ms": 289.7062080446631, "ttft_ms": 13.426082907244563, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "8af6edf6736154e1eb1c85c2253cd550117bde6765534baeea96c12d635d74d1", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "61f91b63aa5b27ddf1638f29", "attempt_id": "30e408db4981", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:52:16.363306Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 1125, "message_tokens": 1037, "message_utf8_bytes": 2388, "cache_tensor_bytes": null, "serialized_payload_bytes": 2388, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13968.575833016075, "ttft_ms": 13.326500076800585, "decode_ms": 119.18487504590303, "receiver_prefill_ms": 121.47708400152624}, "total_wall_ms": 14224.431583075784, "ttft_ms": 13.326500076800585, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "9,220,745", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "001f9d0e8804c7203a8dd550d4e1adc5e0abcc0cc43f59ed18e0373f350318f3", "output_hash": "ae55ba88b2cb1b0f1a4e0960ae312e0a1c177c79a6d14b67bff1677cec19b649"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "42400a4e08681bf6d2e32436", "attempt_id": "0f12353d19ca", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:52:16.634330Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 1127, "message_tokens": 1037, "message_utf8_bytes": 2388, "cache_tensor_bytes": null, "serialized_payload_bytes": 2388, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13968.575833016075, "ttft_ms": 13.3563750423491, "decode_ms": 132.6133330585435, "receiver_prefill_ms": 122.07945808768272}, "total_wall_ms": 14238.473541103303, "ttft_ms": 13.3563750423491, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "LY-51E-3199", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "001f9d0e8804c7203a8dd550d4e1adc5e0abcc0cc43f59ed18e0373f350318f3", "output_hash": "a1245d1536890a7e4b026366df48441d503f491f782cfd388797b5782443c5c7"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b133f3a53367f3fa53f411b3", "attempt_id": "8d5da6859fea", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:52:16.919469Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 1131, "message_tokens": 1037, "message_utf8_bytes": 2388, "cache_tensor_bytes": null, "serialized_payload_bytes": 2388, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13968.575833016075, "ttft_ms": 13.434458989650011, "decode_ms": 93.32300000824034, "receiver_prefill_ms": 121.81470799259841}, "total_wall_ms": 14198.863832978532, "ttft_ms": 13.434458989650011, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "2286335", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "001f9d0e8804c7203a8dd550d4e1adc5e0abcc0cc43f59ed18e0373f350318f3", "output_hash": "f937964a11b672d870cf32a715aa3148a19032710b9c33607b5654b9b8e34945"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "50f59e3d08beb98e703c0317", "attempt_id": "e6cfb442cf1d", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:52:17.164560Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 1127, "message_tokens": 1037, "message_utf8_bytes": 2388, "cache_tensor_bytes": null, "serialized_payload_bytes": 2388, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13968.575833016075, "ttft_ms": 13.776957988739014, "decode_ms": 106.44770797807723, "receiver_prefill_ms": 120.88895891793072}, "total_wall_ms": 14211.414541117847, "ttft_ms": 13.776957988739014, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "10440170", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "001f9d0e8804c7203a8dd550d4e1adc5e0abcc0cc43f59ed18e0373f350318f3", "output_hash": "e4ac77f90dc0972fc3e3badd118c283c5cd2781a19aef45cabd84dbebaed28f6"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "9ac36a38421afec5fe87ae72", "attempt_id": "fb9f7ce368ff", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:52:17.422279Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 1133, "message_tokens": 1037, "message_utf8_bytes": 2388, "cache_tensor_bytes": null, "serialized_payload_bytes": 2388, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13968.575833016075, "ttft_ms": 13.102625031024218, "decode_ms": 66.12383294850588, "receiver_prefill_ms": 121.16787489503622}, "total_wall_ms": 14170.66545796115, "ttft_ms": 13.102625031024218, "input_usage": {}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Wendra-7", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "001f9d0e8804c7203a8dd550d4e1adc5e0abcc0cc43f59ed18e0373f350318f3", "output_hash": "bd81b870addca9eb7c7a0adcc4951f2aba6a86d8408adc5d88e9806bf8dd99e5"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "aa86ac0f8e543d5a53b894a6", "attempt_id": "a40b096c478e", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:52:17.638770Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 1128, "message_tokens": 1037, "message_utf8_bytes": 2388, "cache_tensor_bytes": null, "serialized_payload_bytes": 2388, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13968.575833016075, "ttft_ms": 13.812834047712386, "decode_ms": 120.08120794780552, "receiver_prefill_ms": 120.94125000294298}, "total_wall_ms": 14225.14591703657, "ttft_ms": 13.812834047712386, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "6,191,848", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "001f9d0e8804c7203a8dd550d4e1adc5e0abcc0cc43f59ed18e0373f350318f3", "output_hash": "1119f4e61c20051b5134c339d65031ec5e2df9ee64d464461b613f5127313c39"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "ae6177619c3d24b1cdcf2274", "attempt_id": "b831db40f148", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:52:17.910500Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 1128, "message_tokens": 1037, "message_utf8_bytes": 2388, "cache_tensor_bytes": null, "serialized_payload_bytes": 2388, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13968.575833016075, "ttft_ms": 13.289791997522116, "decode_ms": 13.077583047561347, "receiver_prefill_ms": 121.3895829860121}, "total_wall_ms": 14118.097207974643, "ttft_ms": 13.289791997522116, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "001f9d0e8804c7203a8dd550d4e1adc5e0abcc0cc43f59ed18e0373f350318f3", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "17a94e51a38fcbc1acd95d85", "attempt_id": "e6097d100fd6", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:52:18.075576Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 1130, "message_tokens": 1037, "message_utf8_bytes": 2388, "cache_tensor_bytes": null, "serialized_payload_bytes": 2388, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13968.575833016075, "ttft_ms": 13.277458027005196, "decode_ms": 13.004167005419731, "receiver_prefill_ms": 121.15975003689528}, "total_wall_ms": 14117.69079102669, "ttft_ms": 13.277458027005196, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "001f9d0e8804c7203a8dd550d4e1adc5e0abcc0cc43f59ed18e0373f350318f3", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "97feea3ace8398d35c425e4f", "attempt_id": "7d55a81c3da0", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:52:32.213273Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 15}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 686, "message_tokens": 595, "message_utf8_bytes": 1540, "cache_tensor_bytes": null, "serialized_payload_bytes": 1540, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13973.957874928601, "ttft_ms": 17.048750072717667, "decode_ms": 92.73666702210903, "receiver_prefill_ms": 76.98704197537154}, "total_wall_ms": 14162.104041897692, "ttft_ms": 17.048750072717667, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "9220745", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "2ef337c50fd670371422bb802aafd1216d25be8103263c0ea6e2e02a9b44b39f", "output_hash": "f5d77aa763f6365bc80d01b9e6e23bc43fb570ac45771d7bd628c75cde727b35"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "1aec5f431dc2727ac57c40bd", "attempt_id": "8dfdbbbf2bd0", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:52:32.416460Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 15}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 688, "message_tokens": 595, "message_utf8_bytes": 1540, "cache_tensor_bytes": null, "serialized_payload_bytes": 1540, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13973.957874928601, "ttft_ms": 13.69704189710319, "decode_ms": 131.48950005415827, "receiver_prefill_ms": 76.3399169081822}, "total_wall_ms": 14196.757374913432, "ttft_ms": 13.69704189710319, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "HE-50G-4487", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "2ef337c50fd670371422bb802aafd1216d25be8103263c0ea6e2e02a9b44b39f", "output_hash": "c07ce171b82e8565f51ddfda3b1f5efae293d5aa9f543d848676ac74378bf413"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "3390ed3e0eb0241d0d04bc30", "attempt_id": "e5aefcf56c9e", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:52:32.654217Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 15}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 692, "message_tokens": 595, "message_utf8_bytes": 1540, "cache_tensor_bytes": null, "serialized_payload_bytes": 1540, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13973.957874928601, "ttft_ms": 13.298750040121377, "decode_ms": 119.5659579243511, "receiver_prefill_ms": 76.7557080835104}, "total_wall_ms": 14184.883207897656, "ttft_ms": 13.298750040121377, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "2,705,965", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "2ef337c50fd670371422bb802aafd1216d25be8103263c0ea6e2e02a9b44b39f", "output_hash": "dbedb6f5d00e413d915513b85d8ed195b18b9acf392ece342a85e15e3752bcc8"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "dba12e05af7a14d063da0efc", "attempt_id": "b3fe1330b44e", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:52:32.879941Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 15}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 688, "message_tokens": 595, "message_utf8_bytes": 1540, "cache_tensor_bytes": null, "serialized_payload_bytes": 1540, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13973.957874928601, "ttft_ms": 13.402875047177076, "decode_ms": 105.64633295871317, "receiver_prefill_ms": 76.20333402883261}, "total_wall_ms": 14170.4764169408, "ttft_ms": 13.402875047177076, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "10440170", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "2ef337c50fd670371422bb802aafd1216d25be8103263c0ea6e2e02a9b44b39f", "output_hash": "e4ac77f90dc0972fc3e3badd118c283c5cd2781a19aef45cabd84dbebaed28f6"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a6da2d6c75eed582336d84e4", "attempt_id": "e1fc2db123ee", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:52:33.091786Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 15}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 694, "message_tokens": 595, "message_utf8_bytes": 1540, "cache_tensor_bytes": null, "serialized_payload_bytes": 1540, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13973.957874928601, "ttft_ms": 13.169166049920022, "decode_ms": 53.148459061048925, "receiver_prefill_ms": 76.95258304011077}, "total_wall_ms": 14118.501333869062, "ttft_ms": 13.169166049920022, "input_usage": {}, "output_usage": {"output_tokens": 4}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Wendra-7", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "2ef337c50fd670371422bb802aafd1216d25be8103263c0ea6e2e02a9b44b39f", "output_hash": "bd81b870addca9eb7c7a0adcc4951f2aba6a86d8408adc5d88e9806bf8dd99e5"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "4348744571963eb51ec9e493", "attempt_id": "7da2c34c67ac", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:52:33.251326Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 15}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 689, "message_tokens": 595, "message_utf8_bytes": 1540, "cache_tensor_bytes": null, "serialized_payload_bytes": 1540, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13973.957874928601, "ttft_ms": 13.34062498062849, "decode_ms": 79.00762499775738, "receiver_prefill_ms": 76.11141598317772}, "total_wall_ms": 14143.722499837168, "ttft_ms": 13.34062498062849, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "791765", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "2ef337c50fd670371422bb802aafd1216d25be8103263c0ea6e2e02a9b44b39f", "output_hash": "248a4d75b6b040869fa44487b5f42fb270b5a342ee46bd0fd0562ea8b78b34ae"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "105c0cee624b7adb04f9d7b7", "attempt_id": "149fb9ec18b9", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:52:33.436233Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 15}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 689, "message_tokens": 595, "message_utf8_bytes": 1540, "cache_tensor_bytes": null, "serialized_payload_bytes": 1540, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13973.957874928601, "ttft_ms": 13.012457988224924, "decode_ms": 13.275665929540992, "receiver_prefill_ms": 76.37820800300688}, "total_wall_ms": 14077.90791685693, "ttft_ms": 13.012457988224924, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "2ef337c50fd670371422bb802aafd1216d25be8103263c0ea6e2e02a9b44b39f", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "6eccd1e5cb5a729db6e128b6", "attempt_id": "b9a873eb99f4", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:52:33.555735Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 15}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 691, "message_tokens": 595, "message_utf8_bytes": 1540, "cache_tensor_bytes": null, "serialized_payload_bytes": 1540, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13973.957874928601, "ttft_ms": 13.155333930626512, "decode_ms": 13.146584038622677, "receiver_prefill_ms": 76.41145901288837}, "total_wall_ms": 14078.037624945864, "ttft_ms": 13.155333930626512, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "2ef337c50fd670371422bb802aafd1216d25be8103263c0ea6e2e02a9b44b39f", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "44b9aea19b9cbb120ff25a6b", "attempt_id": "ecd26f8b988e", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:52:47.627347Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 1389, "message_tokens": 1299, "message_utf8_bytes": 3345, "cache_tensor_bytes": null, "serialized_payload_bytes": 3345, "side_information_bytes": 0, "fetched_bytes": 1797, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13952.552167000249, "ttft_ms": 13.434499967843294, "decode_ms": 93.5752079822123, "receiver_prefill_ms": 149.01266701053828}, "total_wall_ms": 14210.848666960374, "ttft_ms": 13.434499967843294, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "9220745", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "94f1a04c6a474709252db6b39eb5fa48348e72d7dbb5cb4ff536e196d41ff683", "output_hash": "f5d77aa763f6365bc80d01b9e6e23bc43fb570ac45771d7bd628c75cde727b35"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "297908bb98493b10c3eef4e9", "attempt_id": "8eb955119eb6", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:52:47.901686Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 1266, "message_tokens": 1174, "message_utf8_bytes": 3032, "cache_tensor_bytes": null, "serialized_payload_bytes": 3032, "side_information_bytes": 0, "fetched_bytes": 1518, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13952.552167000249, "ttft_ms": 13.354958035051823, "decode_ms": 133.4893750026822, "receiver_prefill_ms": 131.136666983366}, "total_wall_ms": 14232.536833034828, "ttft_ms": 13.354958035051823, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "LY-51E-3199", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "63ceb7e1cbf16cb61185ee7e9a2220f5f52a990467c4a0cd2346afcd57b7fff0", "output_hash": "a1245d1536890a7e4b026366df48441d503f491f782cfd388797b5782443c5c7"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "36dd8046515df839e42d0b54", "attempt_id": "b9bca6ebaff9", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:52:48.197041Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 1361, "message_tokens": 1265, "message_utf8_bytes": 3226, "cache_tensor_bytes": null, "serialized_payload_bytes": 3226, "side_information_bytes": 0, "fetched_bytes": 1693, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13952.552167000249, "ttft_ms": 13.199749984778464, "decode_ms": 93.36437494494021, "receiver_prefill_ms": 143.83679104503244}, "total_wall_ms": 14205.0389170181, "ttft_ms": 13.199749984778464, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "2705965", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "542b812936fd86510dea21f0f1e2a2178b6f805f4ea71b4a7f08b7237892cf18", "output_hash": "149792af0b004a6863128878a817e8506bb3a8476a6d59878795e90cba3743a6"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "f748fe570c5f409422a9c54e", "attempt_id": "71a2081ef8a2", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:52:48.464576Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 1381, "message_tokens": 1289, "message_utf8_bytes": 3301, "cache_tensor_bytes": null, "serialized_payload_bytes": 3301, "side_information_bytes": 0, "fetched_bytes": 1758, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13952.552167000249, "ttft_ms": 13.362957979552448, "decode_ms": 93.2094999589026, "receiver_prefill_ms": 147.8106250287965}, "total_wall_ms": 14209.076167084277, "ttft_ms": 13.362957979552448, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "9127965", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "c98f855f6456f226d528750aa7ff38d4cf40e0ac35efa094d789799a21a25ab4", "output_hash": "a0c871941224304cec47bb1b3d64b8706843b299f14ab94899d97df14f433da7"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "864cd800c761fa66d228dbeb", "attempt_id": "8cee25b5d2ba", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:52:48.736183Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 1341, "message_tokens": 1243, "message_utf8_bytes": 3193, "cache_tensor_bytes": null, "serialized_payload_bytes": 3193, "side_information_bytes": 0, "fetched_bytes": 1665, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13952.552167000249, "ttft_ms": 14.310959028080106, "decode_ms": 65.98262500483543, "receiver_prefill_ms": 142.7657080348581}, "total_wall_ms": 14177.674959064461, "ttft_ms": 14.310959028080106, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Wendra-7", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "1107c88b6f0e9fea1bc16d1336d4a02614c15f7de1fc7aecb13b343dd3a61bf3", "output_hash": "bd81b870addca9eb7c7a0adcc4951f2aba6a86d8408adc5d88e9806bf8dd99e5"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "42bc4a904cf5f1a5faa89e36", "attempt_id": "cb5b5c34787f", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:52:48.976372Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 1392, "message_tokens": 1299, "message_utf8_bytes": 3345, "cache_tensor_bytes": null, "serialized_payload_bytes": 3345, "side_information_bytes": 0, "fetched_bytes": 1797, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13952.552167000249, "ttft_ms": 13.382583041675389, "decode_ms": 106.92658403422683, "receiver_prefill_ms": 148.4101670794189}, "total_wall_ms": 14223.410584032536, "ttft_ms": 13.382583041675389, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "24311905", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "43826e15cfde7c5490bde0ec0a6ac4cda3c45598ffb5d907d0753935c9db9e97", "output_hash": "6df9c97086992781d7006a81c02f98c27a9e29e1e1a523a88511299e21cf0ec9"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "0958903e8d94a9610d323a74", "attempt_id": "0c9d1ec142ce", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:52:49.262060Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 1366, "message_tokens": 1273, "message_utf8_bytes": 3286, "cache_tensor_bytes": null, "serialized_payload_bytes": 3286, "side_information_bytes": 0, "fetched_bytes": 1737, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13952.552167000249, "ttft_ms": 13.348375097848475, "decode_ms": 13.512292061932385, "receiver_prefill_ms": 144.59429203998297}, "total_wall_ms": 14126.070333993994, "ttft_ms": 13.348375097848475, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "105d73403992b03ac3f8be72610b60d63ff4f426f7dc29aa5f761afa6919796a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "6c230f0b4bfa599ea3c520cd", "attempt_id": "bab8c3dd6e79", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:52:49.450865Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 1384, "message_tokens": 1289, "message_utf8_bytes": 3301, "cache_tensor_bytes": null, "serialized_payload_bytes": 3301, "side_information_bytes": 0, "fetched_bytes": 1758, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13952.552167000249, "ttft_ms": 13.223416055552661, "decode_ms": 13.20341695100069, "receiver_prefill_ms": 147.87545800209045}, "total_wall_ms": 14128.905583987944, "ttft_ms": 13.223416055552661, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "800ac9f7747c62da981398b530f42248bcb3c8da38fc2efc6f7a1bdd5987349a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c07d0953e5aa037ab958ec62", "attempt_id": "72adf28110d5", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:52:49.898282Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 86040576}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 44, "message_tokens": null, "message_utf8_bytes": 8172, "cache_tensor_bytes": 86040576, "serialized_payload_bytes": 86040576, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 256.0949169564992, "ttft_ms": 28.84633291978389, "decode_ms": 123.63091704901308, "receiver_prefill_ms": 0.0}, "total_wall_ms": 417.922583874315, "ttft_ms": 28.84633291978389, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "9,220,745", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "ae55ba88b2cb1b0f1a4e0960ae312e0a1c177c79a6d14b67bff1677cec19b649"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "882d16fff68d66fc69e5f994", "attempt_id": "d31221ca869f", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:52:50.075224Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 86040576}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 46, "message_tokens": null, "message_utf8_bytes": 8172, "cache_tensor_bytes": 86040576, "serialized_payload_bytes": 86040576, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 256.0949169564992, "ttft_ms": 27.29529095813632, "decode_ms": 138.53908295277506, "receiver_prefill_ms": 0.0}, "total_wall_ms": 425.72837602347136, "ttft_ms": 27.29529095813632, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "LY-51E-3199", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "a1245d1536890a7e4b026366df48441d503f491f782cfd388797b5782443c5c7"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "7949b1f17f556af831b1f920", "attempt_id": "b3e1ce431c3b", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:52:50.259672Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 86040576}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 50, "message_tokens": null, "message_utf8_bytes": 8172, "cache_tensor_bytes": 86040576, "serialized_payload_bytes": 86040576, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 256.0949169564992, "ttft_ms": 27.31866599060595, "decode_ms": 96.75316594075412, "receiver_prefill_ms": 0.0}, "total_wall_ms": 383.0571670550853, "ttft_ms": 27.31866599060595, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "2126290", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "a030b162f5adef1d9151ead9ebe84b7d516e2701c0e6988ecf196d2652bfd94e"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "1222f84ac590c5c17634c517", "attempt_id": "75928ff92f8c", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:52:50.402195Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 86040576}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 46, "message_tokens": null, "message_utf8_bytes": 8172, "cache_tensor_bytes": 86040576, "serialized_payload_bytes": 86040576, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 256.0949169564992, "ttft_ms": 27.03141700476408, "decode_ms": 124.11045795306563, "receiver_prefill_ms": 0.0}, "total_wall_ms": 411.12075094133615, "ttft_ms": 27.03141700476408, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1,034,805", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b1dc6064f97c41a63e2b857664a4196c5397f8224728e25ada16fcf249b7ed60"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "f1731396b0b54ffcdb4b459d", "attempt_id": "9a512b063b8d", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:52:50.572418Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 86040576}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 52, "message_tokens": null, "message_utf8_bytes": 8172, "cache_tensor_bytes": 86040576, "serialized_payload_bytes": 86040576, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 256.0949169564992, "ttft_ms": 28.035875060595572, "decode_ms": 69.20516700483859, "receiver_prefill_ms": 0.0}, "total_wall_ms": 357.2627080138773, "ttft_ms": 28.035875060595572, "input_usage": {}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Wendra-7", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "bd81b870addca9eb7c7a0adcc4951f2aba6a86d8408adc5d88e9806bf8dd99e5"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "12173949a746de002dfdc82a", "attempt_id": "1405c6690764", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:52:50.688728Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 86040576}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 8172, "cache_tensor_bytes": 86040576, "serialized_payload_bytes": 86040576, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 256.0949169564992, "ttft_ms": 28.778583044186234, "decode_ms": 110.20866699982435, "receiver_prefill_ms": 0.0}, "total_wall_ms": 399.0811249241233, "ttft_ms": 28.778583044186234, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "10,65111", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "320606274d92925cb62820afc5c46dda682e94f1815c8b7633c447fed708d971"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "aee8114d299d36044e392d34", "attempt_id": "409da5cfa5ad", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:52:50.846990Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 86040576}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 8172, "cache_tensor_bytes": 86040576, "serialized_payload_bytes": 86040576, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 256.0949169564992, "ttft_ms": 28.648584033362567, "decode_ms": 13.611042057164013, "receiver_prefill_ms": 0.0}, "total_wall_ms": 302.1001670276746, "ttft_ms": 28.648584033362567, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "32de73da3eb10529dfcb5b76", "attempt_id": "92d69a30b017", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:52:50.907842Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 86040576}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 49, "message_tokens": null, "message_utf8_bytes": 8172, "cache_tensor_bytes": 86040576, "serialized_payload_bytes": 86040576, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 256.0949169564992, "ttft_ms": 28.503417037427425, "decode_ms": 13.747834018431604, "receiver_prefill_ms": 0.0}, "total_wall_ms": 302.2424998925999, "ttft_ms": 28.503417037427425, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "1cc38fbfeb4bc94743d2ecf6", "attempt_id": "c66a468ae669", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:52:50.969498Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 2456, "message_tokens": 2370, "message_utf8_bytes": 8702, "cache_tensor_bytes": null, "serialized_payload_bytes": 8702, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.0374580267816782, "ttft_ms": 13.383166980929673, "decode_ms": 121.00166699383408, "receiver_prefill_ms": 267.93816697318107}, "total_wall_ms": 406.61816706415266, "ttft_ms": 13.383166980929673, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "9,220,745", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "db3d8fa34ef5dbf89b0db7e6dc0d9ce7c48cd2f1c6f7cbbcd25529c75ad49aa3", "output_hash": "ae55ba88b2cb1b0f1a4e0960ae312e0a1c177c79a6d14b67bff1677cec19b649"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "733fc1d6c54e1f2628dc3756", "attempt_id": "4299f409b5b3", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:52:51.391432Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 2458, "message_tokens": 2370, "message_utf8_bytes": 8702, "cache_tensor_bytes": null, "serialized_payload_bytes": 8702, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.0374580267816782, "ttft_ms": 13.504624948836863, "decode_ms": 134.00829094462097, "receiver_prefill_ms": 268.21616699453443}, "total_wall_ms": 419.439708115533, "ttft_ms": 13.504624948836863, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "LY-51E-3199", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "db3d8fa34ef5dbf89b0db7e6dc0d9ce7c48cd2f1c6f7cbbcd25529c75ad49aa3", "output_hash": "a1245d1536890a7e4b026366df48441d503f491f782cfd388797b5782443c5c7"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "4dd868ae3ff53f20905e77f2", "attempt_id": "cccf16327f14", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:52:51.825835Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 2462, "message_tokens": 2370, "message_utf8_bytes": 8702, "cache_tensor_bytes": null, "serialized_payload_bytes": 8702, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.0374580267816782, "ttft_ms": 13.603165978565812, "decode_ms": 121.51316704694182, "receiver_prefill_ms": 268.32029095385224}, "total_wall_ms": 407.2520830668509, "ttft_ms": 13.603165978565812, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "2,705,965", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "db3d8fa34ef5dbf89b0db7e6dc0d9ce7c48cd2f1c6f7cbbcd25529c75ad49aa3", "output_hash": "dbedb6f5d00e413d915513b85d8ed195b18b9acf392ece342a85e15e3752bcc8"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "89fe8dbce5164b02aa52d297", "attempt_id": "b4ca07aeed90", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:52:52.248260Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 2458, "message_tokens": 2370, "message_utf8_bytes": 8702, "cache_tensor_bytes": null, "serialized_payload_bytes": 8702, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.0374580267816782, "ttft_ms": 13.84087500628084, "decode_ms": 121.85299990233034, "receiver_prefill_ms": 266.83970796875656}, "total_wall_ms": 406.3759580021724, "ttft_ms": 13.84087500628084, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1,034,805", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "db3d8fa34ef5dbf89b0db7e6dc0d9ce7c48cd2f1c6f7cbbcd25529c75ad49aa3", "output_hash": "b1dc6064f97c41a63e2b857664a4196c5397f8224728e25ada16fcf249b7ed60"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "34e30f8d61f28618b9a7f084", "attempt_id": "0b3bd1159139", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:52:52.669978Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 2464, "message_tokens": 2370, "message_utf8_bytes": 8702, "cache_tensor_bytes": null, "serialized_payload_bytes": 8702, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.0374580267816782, "ttft_ms": 13.514499994926155, "decode_ms": 67.66441697254777, "receiver_prefill_ms": 267.95337500516325}, "total_wall_ms": 352.8682490577921, "ttft_ms": 13.514499994926155, "input_usage": {}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Felaldnak-7", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "db3d8fa34ef5dbf89b0db7e6dc0d9ce7c48cd2f1c6f7cbbcd25529c75ad49aa3", "output_hash": "2b9a4cf7221f1b6f31508a6b3078ec521f211026a932da5020cddd512ecc5fa2"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "1b410be407de10d479f5904a", "attempt_id": "7b1aabbd9b68", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:52:53.038033Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 2459, "message_tokens": 2370, "message_utf8_bytes": 8702, "cache_tensor_bytes": null, "serialized_payload_bytes": 8702, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.0374580267816782, "ttft_ms": 13.337666052393615, "decode_ms": 134.67466598376632, "receiver_prefill_ms": 268.34308297839016}, "total_wall_ms": 420.0379579560831, "ttft_ms": 13.337666052393615, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "30,478,733", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "db3d8fa34ef5dbf89b0db7e6dc0d9ce7c48cd2f1c6f7cbbcd25529c75ad49aa3", "output_hash": "29d352c0c861aad8cf1cbb1ad2d22919f61918448bf12d3c5395dbebcd75744a"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "240c982e05e3935d6edd342c", "attempt_id": "d000df51bb8e", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:52:53.473724Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 2459, "message_tokens": 2370, "message_utf8_bytes": 8702, "cache_tensor_bytes": null, "serialized_payload_bytes": 8702, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.0374580267816782, "ttft_ms": 13.358791940845549, "decode_ms": 13.535707956179976, "receiver_prefill_ms": 267.95041700825095}, "total_wall_ms": 298.5626660520211, "ttft_ms": 13.358791940845549, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "db3d8fa34ef5dbf89b0db7e6dc0d9ce7c48cd2f1c6f7cbbcd25529c75ad49aa3", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c30dfa6c03242a9bf9fe74f4", "attempt_id": "735b5e12284b", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:52:53.787850Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 2461, "message_tokens": 2370, "message_utf8_bytes": 8702, "cache_tensor_bytes": null, "serialized_payload_bytes": 8702, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.0374580267816782, "ttft_ms": 13.493916019797325, "decode_ms": 13.317500008270144, "receiver_prefill_ms": 267.8430830128491}, "total_wall_ms": 298.2817080337554, "ttft_ms": 13.493916019797325, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "db3d8fa34ef5dbf89b0db7e6dc0d9ce7c48cd2f1c6f7cbbcd25529c75ad49aa3", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "4cf77fc484da9cc969b2039d", "attempt_id": "2b7165842594", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:53:18.301363Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1790}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 1902, "message_tokens": 1814, "message_utf8_bytes": 5707, "cache_tensor_bytes": null, "serialized_payload_bytes": 5707, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24199.87187499646, "ttft_ms": 13.415666064247489, "decode_ms": 120.18529104534537, "receiver_prefill_ms": 202.7512079803273}, "total_wall_ms": 24539.214291027747, "ttft_ms": 13.415666064247489, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "9,220,745", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "5c1ea70cdd3b76f2cd2a8e35362096a3ad57b1255a90dbb8b12fd19103aedc03", "output_hash": "ae55ba88b2cb1b0f1a4e0960ae312e0a1c177c79a6d14b67bff1677cec19b649"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "ec8e76fcaa6ef2fc9d940a5e", "attempt_id": "50afeadfc2f1", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:53:18.656610Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1790}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 1904, "message_tokens": 1814, "message_utf8_bytes": 5707, "cache_tensor_bytes": null, "serialized_payload_bytes": 5707, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24199.87187499646, "ttft_ms": 13.924875063821673, "decode_ms": 133.1768330419436, "receiver_prefill_ms": 202.08233396988362}, "total_wall_ms": 24552.00658296235, "ttft_ms": 13.924875063821673, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "LY-51E-3199", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "5c1ea70cdd3b76f2cd2a8e35362096a3ad57b1255a90dbb8b12fd19103aedc03", "output_hash": "a1245d1536890a7e4b026366df48441d503f491f782cfd388797b5782443c5c7"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "85dc1f232bf2cd1b6144fc58", "attempt_id": "354e599ef034", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:53:19.025698Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1790}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 1908, "message_tokens": 1814, "message_utf8_bytes": 5707, "cache_tensor_bytes": null, "serialized_payload_bytes": 5707, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24199.87187499646, "ttft_ms": 13.520583976060152, "decode_ms": 94.16624996811152, "receiver_prefill_ms": 202.5150420377031}, "total_wall_ms": 24512.920375098474, "ttft_ms": 13.520583976060152, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "2126290", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "5c1ea70cdd3b76f2cd2a8e35362096a3ad57b1255a90dbb8b12fd19103aedc03", "output_hash": "a030b162f5adef1d9151ead9ebe84b7d516e2701c0e6988ecf196d2652bfd94e"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "0ef559f021a8e8efe0080188", "attempt_id": "0753fb72b050", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:53:19.354004Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1790}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 1904, "message_tokens": 1814, "message_utf8_bytes": 5707, "cache_tensor_bytes": null, "serialized_payload_bytes": 5707, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24199.87187499646, "ttft_ms": 13.820249936543405, "decode_ms": 120.9138750564307, "receiver_prefill_ms": 200.68891695700586}, "total_wall_ms": 24538.138292031363, "ttft_ms": 13.820249936543405, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1,034,805", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "5c1ea70cdd3b76f2cd2a8e35362096a3ad57b1255a90dbb8b12fd19103aedc03", "output_hash": "b1dc6064f97c41a63e2b857664a4196c5397f8224728e25ada16fcf249b7ed60"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "28840980a60fe1846bad2075", "attempt_id": "481436f1d33e", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:53:19.707746Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1790}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 1910, "message_tokens": 1814, "message_utf8_bytes": 5707, "cache_tensor_bytes": null, "serialized_payload_bytes": 5707, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24199.87187499646, "ttft_ms": 13.410291983745992, "decode_ms": 66.85479206498712, "receiver_prefill_ms": 202.5980829494074}, "total_wall_ms": 24485.629125032574, "ttft_ms": 13.410291983745992, "input_usage": {}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Wendra-7", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "5c1ea70cdd3b76f2cd2a8e35362096a3ad57b1255a90dbb8b12fd19103aedc03", "output_hash": "bd81b870addca9eb7c7a0adcc4951f2aba6a86d8408adc5d88e9806bf8dd99e5"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "fdad78570273fb50579346f0", "attempt_id": "86ec098a2bf5", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:53:20.009954Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1790}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 1905, "message_tokens": 1814, "message_utf8_bytes": 5707, "cache_tensor_bytes": null, "serialized_payload_bytes": 5707, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24199.87187499646, "ttft_ms": 13.609457993879914, "decode_ms": 94.46420904714614, "receiver_prefill_ms": 201.1519999941811}, "total_wall_ms": 24512.021417031065, "ttft_ms": 13.609457993879914, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1843015", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "5c1ea70cdd3b76f2cd2a8e35362096a3ad57b1255a90dbb8b12fd19103aedc03", "output_hash": "521b81b0cf6f56079ee7d259950cecc49558a3b7f0bbbf791cc9bdf61621c2f5"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "47bf5dfd2cc32de31b424295", "attempt_id": "cea5d0b0fc76", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:53:20.337567Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1790}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 1905, "message_tokens": 1814, "message_utf8_bytes": 5707, "cache_tensor_bytes": null, "serialized_payload_bytes": 5707, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24199.87187499646, "ttft_ms": 13.333291048184037, "decode_ms": 13.470083009451628, "receiver_prefill_ms": 201.41441596206278}, "total_wall_ms": 24431.12404202111, "ttft_ms": 13.333291048184037, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "5c1ea70cdd3b76f2cd2a8e35362096a3ad57b1255a90dbb8b12fd19103aedc03", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "4b9a3c5a39ba2b78529596a5", "attempt_id": "6cd5423b9ac0", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:53:20.584581Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1790}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 1907, "message_tokens": 1814, "message_utf8_bytes": 5707, "cache_tensor_bytes": null, "serialized_payload_bytes": 5707, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24199.87187499646, "ttft_ms": 13.522625085897744, "decode_ms": 13.189875055104494, "receiver_prefill_ms": 200.91699995100498}, "total_wall_ms": 24430.380500038154, "ttft_ms": 13.522625085897744, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "5c1ea70cdd3b76f2cd2a8e35362096a3ad57b1255a90dbb8b12fd19103aedc03", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "2f662c07b703be88c6106914", "attempt_id": "6674ffe5a923", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:53:48.556614Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 30}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 1265, "message_tokens": 1174, "message_utf8_bytes": 3208, "cache_tensor_bytes": null, "serialized_payload_bytes": 3208, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27725.90354201384, "ttft_ms": 13.46270798239857, "decode_ms": 120.47804205212742, "receiver_prefill_ms": 131.05395797174424}, "total_wall_ms": 27992.921500001103, "ttft_ms": 13.46270798239857, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "9,220,745", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "26efab12b854e7bef57db7d21341022cd02a1346e9865344f4e46aedd8c34801", "output_hash": "ae55ba88b2cb1b0f1a4e0960ae312e0a1c177c79a6d14b67bff1677cec19b649"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "e3bc1969108df86e75e4f2bb", "attempt_id": "731b2e6b233a", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:53:48.839558Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 30}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 1267, "message_tokens": 1174, "message_utf8_bytes": 3208, "cache_tensor_bytes": null, "serialized_payload_bytes": 3208, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27725.90354201384, "ttft_ms": 13.334124931134284, "decode_ms": 133.46787495538592, "receiver_prefill_ms": 130.51837508101016}, "total_wall_ms": 28005.161500070244, "ttft_ms": 13.334124931134284, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "LY-51E-3199", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "26efab12b854e7bef57db7d21341022cd02a1346e9865344f4e46aedd8c34801", "output_hash": "a1245d1536890a7e4b026366df48441d503f491f782cfd388797b5782443c5c7"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "cf83faff27cd4fdb6e6c64ed", "attempt_id": "8b056f30c536", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:53:49.134034Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 30}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 1271, "message_tokens": 1174, "message_utf8_bytes": 3208, "cache_tensor_bytes": null, "serialized_payload_bytes": 3208, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27725.90354201384, "ttft_ms": 13.57920910231769, "decode_ms": 120.10275002103299, "receiver_prefill_ms": 131.52591697871685}, "total_wall_ms": 27993.053041980602, "ttft_ms": 13.57920910231769, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "2,705,965", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "26efab12b854e7bef57db7d21341022cd02a1346e9865344f4e46aedd8c34801", "output_hash": "dbedb6f5d00e413d915513b85d8ed195b18b9acf392ece342a85e15e3752bcc8"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b13bd49fe0b74a9bb84aab06", "attempt_id": "7cbc690329bf", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:53:49.417594Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 30}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 1267, "message_tokens": 1174, "message_utf8_bytes": 3208, "cache_tensor_bytes": null, "serialized_payload_bytes": 3208, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27725.90354201384, "ttft_ms": 13.329750043340027, "decode_ms": 120.53437496069819, "receiver_prefill_ms": 130.55154192261398}, "total_wall_ms": 27992.214582976885, "ttft_ms": 13.329750043340027, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1,402,360", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "26efab12b854e7bef57db7d21341022cd02a1346e9865344f4e46aedd8c34801", "output_hash": "c36829f2e67ba37c9a6a3a2f2b3f1c014b7b943848f81fda5d9b866cd0f3f869"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "20a3c6f55022abcdd000cd73", "attempt_id": "f8df35533c7a", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:53:49.700544Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 30}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 1273, "message_tokens": 1174, "message_utf8_bytes": 3208, "cache_tensor_bytes": null, "serialized_payload_bytes": 3208, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27725.90354201384, "ttft_ms": 13.230292010121047, "decode_ms": 67.25029100198299, "receiver_prefill_ms": 831.6094999900088}, "total_wall_ms": 28640.23354195524, "ttft_ms": 13.230292010121047, "input_usage": {}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Felaldnak-7", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "26efab12b854e7bef57db7d21341022cd02a1346e9865344f4e46aedd8c34801", "output_hash": "2b9a4cf7221f1b6f31508a6b3078ec521f211026a932da5020cddd512ecc5fa2"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "7d2050a7b72c131d31d236c4", "attempt_id": "7fab8d1e7433", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:53:50.629868Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 30}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 1268, "message_tokens": 1174, "message_utf8_bytes": 3208, "cache_tensor_bytes": null, "serialized_payload_bytes": 3208, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27725.90354201384, "ttft_ms": 14.136582962237298, "decode_ms": 99.7018750058487, "receiver_prefill_ms": 738.3912079967558}, "total_wall_ms": 28580.18941699993, "ttft_ms": 14.136582962237298, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1634145", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "26efab12b854e7bef57db7d21341022cd02a1346e9865344f4e46aedd8c34801", "output_hash": "c2efea663bc150cda4d360a3f86d3882c4898576a8a1f16a4fd1eaa7b12f3b5a"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "30b1e10e83f90d3cddea2909", "attempt_id": "a2ac15da195a", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:53:51.499499Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 30}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 1268, "message_tokens": 1174, "message_utf8_bytes": 3208, "cache_tensor_bytes": null, "serialized_payload_bytes": 3208, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27725.90354201384, "ttft_ms": 13.37237493135035, "decode_ms": 13.2672080071643, "receiver_prefill_ms": 131.63125002756715}, "total_wall_ms": 27886.517625069246, "ttft_ms": 13.37237493135035, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "26efab12b854e7bef57db7d21341022cd02a1346e9865344f4e46aedd8c34801", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c318bc138bebd7f02a4260cb", "attempt_id": "d7271a7d2e8a", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:53:51.678410Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 30}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 1270, "message_tokens": 1174, "message_utf8_bytes": 3208, "cache_tensor_bytes": null, "serialized_payload_bytes": 3208, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27725.90354201384, "ttft_ms": 15.089042019098997, "decode_ms": 13.995541026815772, "receiver_prefill_ms": 165.11958301998675}, "total_wall_ms": 27922.26958402898, "ttft_ms": 15.089042019098997, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "26efab12b854e7bef57db7d21341022cd02a1346e9865344f4e46aedd8c34801", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "953380b8c45f7ef4db2137e2", "attempt_id": "04beba741d30", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:54:19.791313Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 2667, "message_tokens": 2577, "message_utf8_bytes": 6912, "cache_tensor_bytes": null, "serialized_payload_bytes": 6912, "side_information_bytes": 0, "fetched_bytes": 3469, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27901.69883298222, "ttft_ms": 14.39045905135572, "decode_ms": 98.4360829461366, "receiver_prefill_ms": 450.4917919402942}, "total_wall_ms": 28472.12845797185, "ttft_ms": 14.39045905135572, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "9220745", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "18b7ec1a031270d9f10208b94558e4102e8280535313d0c97b3b71f314005c78", "output_hash": "f5d77aa763f6365bc80d01b9e6e23bc43fb570ac45771d7bd628c75cde727b35"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "56f821ad82703fa50867625c", "attempt_id": "11ae5b2ab56e", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:54:20.383526Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 2473, "message_tokens": 2381, "message_utf8_bytes": 6580, "cache_tensor_bytes": null, "serialized_payload_bytes": 6580, "side_information_bytes": 0, "fetched_bytes": 3145, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27901.69883298222, "ttft_ms": 13.988499995321035, "decode_ms": 134.00245807133615, "receiver_prefill_ms": 270.0579169904813}, "total_wall_ms": 28326.97333302349, "ttft_ms": 13.988499995321035, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "LY-51E-3199", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "f577e52e006359ba95a635cb4a7662c4a98a57cda6d0c2378bf66d5ee98c9111", "output_hash": "a1245d1536890a7e4b026366df48441d503f491f782cfd388797b5782443c5c7"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "70d922dbdb6ae609c998c326", "attempt_id": "53a65685ee77", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:54:20.824523Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 2691, "message_tokens": 2595, "message_utf8_bytes": 6872, "cache_tensor_bytes": null, "serialized_payload_bytes": 6872, "side_information_bytes": 0, "fetched_bytes": 3437, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27901.69883298222, "ttft_ms": 13.310958980582654, "decode_ms": 93.63141609355807, "receiver_prefill_ms": 322.5940829142928}, "total_wall_ms": 28334.70420795493, "ttft_ms": 13.310958980582654, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "2705965", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "5a28ff96ff2c68230de8211edca6f6412808435f8445c25cc176cebdccda2695", "output_hash": "149792af0b004a6863128878a817e8506bb3a8476a6d59878795e90cba3743a6"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b98f68cf46bc9327471db010", "attempt_id": "67722319fb64", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:54:21.272099Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 2667, "message_tokens": 2575, "message_utf8_bytes": 6845, "cache_tensor_bytes": null, "serialized_payload_bytes": 6845, "side_information_bytes": 0, "fetched_bytes": 3417, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27901.69883298222, "ttft_ms": 13.565375003963709, "decode_ms": 95.09349998552352, "receiver_prefill_ms": 288.91737502999604}, "total_wall_ms": 28302.76816606056, "ttft_ms": 13.565375003963709, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1034805", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "4350ddeccbd383bcd6dfbdbb9bb58abf18cebe1e7a85cd75ae5e47bb7af836f0", "output_hash": "c425daa91055a677e1c93ecb92f96e4b1e9019e07a9174949e04bad9244e64a5"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b8f5f57186a4024c14a6d8f6", "attempt_id": "8a134927200a", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:54:21.687566Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 2576, "message_tokens": 2478, "message_utf8_bytes": 6732, "cache_tensor_bytes": null, "serialized_payload_bytes": 6732, "side_information_bytes": 0, "fetched_bytes": 3312, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27901.69883298222, "ttft_ms": 13.481374946422875, "decode_ms": 66.73941703047603, "receiver_prefill_ms": 302.4657090427354}, "total_wall_ms": 28288.003749912605, "ttft_ms": 13.481374946422875, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Felaldnak-7", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "89a18e3ce6e51234c3b6d1ea069c2f34c0fd16c85179a7a1f5fd202b4f614f40", "output_hash": "2b9a4cf7221f1b6f31508a6b3078ec521f211026a932da5020cddd512ecc5fa2"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "d7e2d1647034273f90608f34", "attempt_id": "62de06b0a124", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:54:22.088268Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 2688, "message_tokens": 2595, "message_utf8_bytes": 6872, "cache_tensor_bytes": null, "serialized_payload_bytes": 6872, "side_information_bytes": 0, "fetched_bytes": 3437, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27901.69883298222, "ttft_ms": 14.06483305618167, "decode_ms": 94.79495894629508, "receiver_prefill_ms": 290.9614999080077}, "total_wall_ms": 28305.022249929607, "ttft_ms": 14.06483305618167, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1463105", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "68e2e189a8ababdd5c2b92c6f791e0b086d1e08bf16e8a361d81e1b2e72774aa", "output_hash": "414937ed43c61fe7d54fbff0ca2c136407fda9e4a235e18d933b5d37548941b5"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "93f104cfd9062caf220edf41", "attempt_id": "996fd38f7266", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:54:22.505859Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 2621, "message_tokens": 2528, "message_utf8_bytes": 6814, "cache_tensor_bytes": null, "serialized_payload_bytes": 6814, "side_information_bytes": 0, "fetched_bytes": 3350, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27901.69883298222, "ttft_ms": 13.610999914817512, "decode_ms": 13.24012503027916, "receiver_prefill_ms": 283.7691659806296}, "total_wall_ms": 28216.038250015117, "ttft_ms": 13.610999914817512, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "e30b710c457dbb7089c184d8cf5cd875104e00edc13665545312cf324319525a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "ebce6f2057b35a34fd73b3f3", "attempt_id": "c505fa5e1f15", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:54:22.835620Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 2654, "message_tokens": 2559, "message_utf8_bytes": 6872, "cache_tensor_bytes": null, "serialized_payload_bytes": 6872, "side_information_bytes": 0, "fetched_bytes": 3435, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27901.69883298222, "ttft_ms": 14.026333927176893, "decode_ms": 13.626708067022264, "receiver_prefill_ms": 344.2983750719577}, "total_wall_ms": 28277.29908307083, "ttft_ms": 14.026333927176893, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "91495f760418b34bbf6182feae7d3ac2ef5b1d92421042244d032562b655d730", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "52cb8f6cac44cd911bf91cdd", "attempt_id": "0cf507c210ae", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:54:23.878243Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 93118464}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 44, "message_tokens": null, "message_utf8_bytes": 8648, "cache_tensor_bytes": 93118464, "serialized_payload_bytes": 93118464, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 652.099582948722, "ttft_ms": 30.06362495943904, "decode_ms": 125.90049998834729, "receiver_prefill_ms": 0.0}, "total_wall_ms": 811.9812490185723, "ttft_ms": 30.06362495943904, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "2,542,075", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "96f4b331f1f468ea1c8e321548942064cd60be1b4c08ae4520eb054e54407062"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c2a6ca1ddff285508c59f000", "attempt_id": "119c8596a60b", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:54:24.055507Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 93118464}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 46, "message_tokens": null, "message_utf8_bytes": 8648, "cache_tensor_bytes": 93118464, "serialized_payload_bytes": 93118464, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 652.099582948722, "ttft_ms": 155.65083408728242, "decode_ms": 138.84158397559077, "receiver_prefill_ms": 0.0}, "total_wall_ms": 949.9572908971459, "ttft_ms": 155.65083408728242, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "LY-51E-3199", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "a1245d1536890a7e4b026366df48441d503f491f782cfd388797b5782443c5c7"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b5d09c2864de553f3eb5ff1e", "attempt_id": "93a61c5215b6", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:54:24.368834Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 93118464}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 50, "message_tokens": null, "message_utf8_bytes": 8648, "cache_tensor_bytes": 93118464, "serialized_payload_bytes": 93118464, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 652.099582948722, "ttft_ms": 161.45566606428474, "decode_ms": 111.65737500414252, "receiver_prefill_ms": 0.0}, "total_wall_ms": 928.0916239367798, "ttft_ms": 161.45566606428474, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "10440170", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "e4ac77f90dc0972fc3e3badd118c283c5cd2781a19aef45cabd84dbebaed28f6"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "7d7bbd562c68d0661097f83f", "attempt_id": "6cf404fda2cf", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:54:24.660334Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 93118464}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 46, "message_tokens": null, "message_utf8_bytes": 8648, "cache_tensor_bytes": 93118464, "serialized_payload_bytes": 93118464, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 652.099582948722, "ttft_ms": 27.87229197565466, "decode_ms": 124.67754096724093, "receiver_prefill_ms": 0.0}, "total_wall_ms": 807.7325829071924, "ttft_ms": 27.87229197565466, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1,034,805", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b1dc6064f97c41a63e2b857664a4196c5397f8224728e25ada16fcf249b7ed60"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "5cae2d5cd77d97008b00ae11", "attempt_id": "f139a9c4099b", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:54:24.831286Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 93118464}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 52, "message_tokens": null, "message_utf8_bytes": 8648, "cache_tensor_bytes": 93118464, "serialized_payload_bytes": 93118464, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 652.099582948722, "ttft_ms": 64.2810839926824, "decode_ms": 69.65558300726116, "receiver_prefill_ms": 0.0}, "total_wall_ms": 789.1524579608813, "ttft_ms": 64.2810839926824, "input_usage": {}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Felaldnak-7", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "2b9a4cf7221f1b6f31508a6b3078ec521f211026a932da5020cddd512ecc5fa2"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "1148d77759a6809205ad9f6b", "attempt_id": "70b13768e008", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:54:24.983287Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 93118464}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 8648, "cache_tensor_bytes": 93118464, "serialized_payload_bytes": 93118464, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 652.099582948722, "ttft_ms": 55.07104191929102, "decode_ms": 111.66929092723876, "receiver_prefill_ms": 0.0}, "total_wall_ms": 822.0470410306007, "ttft_ms": 55.07104191929102, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "18224243", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "7c929fa409303ce767bf7555729f41dcad2f8d73d36a7ba3fca2961ca9ef1184"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b3f21ed8db272e664a145234", "attempt_id": "6a7d4bcff741", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:54:25.168786Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 93118464}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 8648, "cache_tensor_bytes": 93118464, "serialized_payload_bytes": 93118464, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 652.099582948722, "ttft_ms": 30.07004107348621, "decode_ms": 14.060416957363486, "receiver_prefill_ms": 0.0}, "total_wall_ms": 699.1643749643117, "ttft_ms": 30.07004107348621, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "f733bd20a45ea1e581bb7bfe", "attempt_id": "1840ad70177b", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:54:25.231721Z", "task_family": "synthetic", "document_id": "syn_03000_2000", "document_hash": "c9cc27c20b05b1f6e592846e84990ee9b69c42f9720a8e338ec2f6fca777fdbc", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 93118464}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1904, "query_tokens": 49, "message_tokens": null, "message_utf8_bytes": 8648, "cache_tensor_bytes": 93118464, "serialized_payload_bytes": 93118464, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 652.099582948722, "ttft_ms": 29.50662502553314, "decode_ms": 13.97612493019551, "receiver_prefill_ms": 0.0}, "total_wall_ms": 698.4850829467177, "ttft_ms": 29.50662502553314, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lunmel-9", "bogus_amount": "2,126,290", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03000_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b71fc065de1e68671a604e9d", "attempt_id": "346ff6beeaca", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:54:26.272214Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 6199, "message_tokens": 6111, "message_utf8_bytes": 30206, "cache_tensor_bytes": null, "serialized_payload_bytes": 30206, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.9838750120252371, "ttft_ms": 13.630458037368953, "decode_ms": 123.25299996882677, "receiver_prefill_ms": 734.6355830086395}, "total_wall_ms": 882.8869590070099, "ttft_ms": 13.630458037368953, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5,755,305", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "d6f0908314091536b1e42844fd9fc7f05ad41be53c6c8a0e550db4825a041c4d", "output_hash": "2feb744bc0efb44064e0bc375d54165760453fdd2c5b159d7c1c555e36e65a43"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "9436541fe45124e8206775e2", "attempt_id": "94c27610a9ae", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:54:27.171570Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 6200, "message_tokens": 6111, "message_utf8_bytes": 30206, "cache_tensor_bytes": null, "serialized_payload_bytes": 30206, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.9838750120252371, "ttft_ms": 14.025792013853788, "decode_ms": 139.4908749498427, "receiver_prefill_ms": 722.2719590645283}, "total_wall_ms": 886.4397919969633, "ttft_ms": 14.025792013853788, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "GY-13E-6844", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "d6f0908314091536b1e42844fd9fc7f05ad41be53c6c8a0e550db4825a041c4d", "output_hash": "2ca6ddf54be9e197af2f12973eddba19b3d0c78d0073a82218a23e1d9bbad92e"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "3b50ab6d641c4a857ba2cf11", "attempt_id": "b9e726f0e6da", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:54:28.074711Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 6204, "message_tokens": 6111, "message_utf8_bytes": 30206, "cache_tensor_bytes": null, "serialized_payload_bytes": 30206, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.9838750120252371, "ttft_ms": 13.739666086621583, "decode_ms": 122.77395790442824, "receiver_prefill_ms": 722.7064999751747}, "total_wall_ms": 869.7892080526799, "ttft_ms": 13.739666086621583, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5,588,115", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "d6f0908314091536b1e42844fd9fc7f05ad41be53c6c8a0e550db4825a041c4d", "output_hash": "6e5b387355d2edb1ffd589b58491594fa792b2b6e99363b2e5710b42eb6429b0"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "3694f2792a4d5883e374ff52", "attempt_id": "d5271f516d22", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:54:28.960925Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 6197, "message_tokens": 6111, "message_utf8_bytes": 30206, "cache_tensor_bytes": null, "serialized_payload_bytes": 30206, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.9838750120252371, "ttft_ms": 13.78758403006941, "decode_ms": 95.72716697584838, "receiver_prefill_ms": 721.7484579887241}, "total_wall_ms": 841.869791969657, "ttft_ms": 13.78758403006941, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4439515", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "d6f0908314091536b1e42844fd9fc7f05ad41be53c6c8a0e550db4825a041c4d", "output_hash": "428041758b16a119eefbcb06d172852931ea08ef516d1c13ef1c0ae220e36a5a"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "bec5fb1affbee5cfa12885af", "attempt_id": "b87888ee8ac7", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:54:29.819476Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 6206, "message_tokens": 6111, "message_utf8_bytes": 30206, "cache_tensor_bytes": null, "serialized_payload_bytes": 30206, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.9838750120252371, "ttft_ms": 13.719749986194074, "decode_ms": 68.17266694270074, "receiver_prefill_ms": 725.5532089620829}, "total_wall_ms": 818.1317909620702, "ttft_ms": 13.719749986194074, "input_usage": {}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Banmunruk-6", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "d6f0908314091536b1e42844fd9fc7f05ad41be53c6c8a0e550db4825a041c4d", "output_hash": "858197b99462f9923666f025810f174515ef1dac3462c948b1ea68e0be358384"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "3b666291e33e7e0d2caba959", "attempt_id": "35118989cb4a", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:54:30.653807Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 6200, "message_tokens": 6111, "message_utf8_bytes": 30206, "cache_tensor_bytes": null, "serialized_payload_bytes": 30206, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.9838750120252371, "ttft_ms": 13.706792029552162, "decode_ms": 109.58924994338304, "receiver_prefill_ms": 727.6161669287831}, "total_wall_ms": 861.5615000016987, "ttft_ms": 13.706792029552162, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "12349929", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "d6f0908314091536b1e42844fd9fc7f05ad41be53c6c8a0e550db4825a041c4d", "output_hash": "488fb449f774b18bf29de75e920d6a58f91248d3f28ad1cb1c028914428618f0"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "7e518fbe643c491c5e07913c", "attempt_id": "b34f5ddd25c1", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:54:31.531468Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 6201, "message_tokens": 6111, "message_utf8_bytes": 30206, "cache_tensor_bytes": null, "serialized_payload_bytes": 30206, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.9838750120252371, "ttft_ms": 14.033292070962489, "decode_ms": 13.938750023953617, "receiver_prefill_ms": 759.1906249290332}, "total_wall_ms": 797.7094169473276, "ttft_ms": 14.033292070962489, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "d6f0908314091536b1e42844fd9fc7f05ad41be53c6c8a0e550db4825a041c4d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "f6758613445e7c4ac85b7421", "attempt_id": "b4ac45710e3d", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:54:32.345745Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 6202, "message_tokens": 6111, "message_utf8_bytes": 30206, "cache_tensor_bytes": null, "serialized_payload_bytes": 30206, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.9838750120252371, "ttft_ms": 13.8352089561522, "decode_ms": 13.510084012523293, "receiver_prefill_ms": 781.313042039983}, "total_wall_ms": 819.362209062092, "ttft_ms": 13.8352089561522, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "d6f0908314091536b1e42844fd9fc7f05ad41be53c6c8a0e550db4825a041c4d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "5f1368ff9bfd73a61e4a190b", "attempt_id": "f598c1574a05", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:54:47.658188Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 1126, "message_tokens": 1037, "message_utf8_bytes": 2602, "cache_tensor_bytes": null, "serialized_payload_bytes": 2602, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14476.044874987565, "ttft_ms": 13.241250067949295, "decode_ms": 93.73670804779977, "receiver_prefill_ms": 121.68554193340242}, "total_wall_ms": 14706.756874918938, "ttft_ms": 13.241250067949295, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5755305", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "557300f625ce5eb2e74d7f0fc79d91bb11361f0872ede0de10ed5ef434f2e0e6", "output_hash": "d62d1e2ffbba79e5e1387defa50c9793b60253cadaf386212b947c0cfb951e11"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b7921d0598be01306e415434", "attempt_id": "1ee78e42500d", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:54:47.907494Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 1127, "message_tokens": 1037, "message_utf8_bytes": 2602, "cache_tensor_bytes": null, "serialized_payload_bytes": 2602, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14476.044874987565, "ttft_ms": 13.261999934911728, "decode_ms": 133.36337497457862, "receiver_prefill_ms": 120.992417098023}, "total_wall_ms": 14745.540334028192, "ttft_ms": 13.261999934911728, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "GY-13E-6844", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "557300f625ce5eb2e74d7f0fc79d91bb11361f0872ede0de10ed5ef434f2e0e6", "output_hash": "2ca6ddf54be9e197af2f12973eddba19b3d0c78d0073a82218a23e1d9bbad92e"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "8afa8f1694324bc16092e31e", "attempt_id": "3472f63e0508", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:54:48.195192Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 1131, "message_tokens": 1037, "message_utf8_bytes": 2602, "cache_tensor_bytes": null, "serialized_payload_bytes": 2602, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14476.044874987565, "ttft_ms": 13.412167085334659, "decode_ms": 93.46608305349946, "receiver_prefill_ms": 121.83425005059689}, "total_wall_ms": 14706.607666914351, "ttft_ms": 13.412167085334659, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5588115", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "557300f625ce5eb2e74d7f0fc79d91bb11361f0872ede0de10ed5ef434f2e0e6", "output_hash": "5e8aa3e317f34247d5545451c63ac9a5b2e481e54d54b520d99a84cd8ee5a8d9"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "e6dcb0095dcf4c9ecbd56ee4", "attempt_id": "c40c68398d5b", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:54:48.443344Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 1124, "message_tokens": 1037, "message_utf8_bytes": 2602, "cache_tensor_bytes": null, "serialized_payload_bytes": 2602, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14476.044874987565, "ttft_ms": 13.435916975140572, "decode_ms": 120.52770901937038, "receiver_prefill_ms": 121.73845898360014}, "total_wall_ms": 14733.606957946904, "ttft_ms": 13.435916975140572, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4,439,515", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "557300f625ce5eb2e74d7f0fc79d91bb11361f0872ede0de10ed5ef434f2e0e6", "output_hash": "7e209a4c9dd6879d7024b31aa971d69ea31077909c7ea20bb771181ec5b64af9"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "096aafc17b808189ac98424a", "attempt_id": "f69452c5a78a", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:54:48.718956Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 1133, "message_tokens": 1037, "message_utf8_bytes": 2602, "cache_tensor_bytes": null, "serialized_payload_bytes": 2602, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14476.044874987565, "ttft_ms": 13.687959057278931, "decode_ms": 66.40074995812029, "receiver_prefill_ms": 121.56941695138812}, "total_wall_ms": 14679.56345807761, "ttft_ms": 13.687959057278931, "input_usage": {}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Banmunruk-6", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "557300f625ce5eb2e74d7f0fc79d91bb11361f0872ede0de10ed5ef434f2e0e6", "output_hash": "858197b99462f9923666f025810f174515ef1dac3462c948b1ea68e0be358384"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "3da3d22bdc26edc25fe60c8d", "attempt_id": "d1f7e78e59d7", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:54:48.940046Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 1127, "message_tokens": 1037, "message_utf8_bytes": 2602, "cache_tensor_bytes": null, "serialized_payload_bytes": 2602, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14476.044874987565, "ttft_ms": 13.408541097305715, "decode_ms": 120.76870899181813, "receiver_prefill_ms": 120.91775005683303}, "total_wall_ms": 14732.993457932025, "ttft_ms": 13.408541097305715, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "18444.485", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "557300f625ce5eb2e74d7f0fc79d91bb11361f0872ede0de10ed5ef434f2e0e6", "output_hash": "a768393110d2a4d6e74ff68b81f633a1096df9c8692303075a4030e0bac16e86"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "01c21c1571874a7445384a95", "attempt_id": "a9d3c7212155", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:54:49.214328Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 1128, "message_tokens": 1037, "message_utf8_bytes": 2602, "cache_tensor_bytes": null, "serialized_payload_bytes": 2602, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14476.044874987565, "ttft_ms": 13.39166599791497, "decode_ms": 13.39316600933671, "receiver_prefill_ms": 121.30500003695488}, "total_wall_ms": 14625.9767919546, "ttft_ms": 13.39166599791497, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "557300f625ce5eb2e74d7f0fc79d91bb11361f0872ede0de10ed5ef434f2e0e6", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "dd79cf3721626b23b3b7c249", "attempt_id": "9effdbd17eba", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:54:49.382071Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 1129, "message_tokens": 1037, "message_utf8_bytes": 2602, "cache_tensor_bytes": null, "serialized_payload_bytes": 2602, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14476.044874987565, "ttft_ms": 13.327874941751361, "decode_ms": 13.337290962226689, "receiver_prefill_ms": 121.58791697584093}, "total_wall_ms": 14626.118000014685, "ttft_ms": 13.327874941751361, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "557300f625ce5eb2e74d7f0fc79d91bb11361f0872ede0de10ed5ef434f2e0e6", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "4eb7f3b9b1697e50ed1ded4b", "attempt_id": "f1dab6f5d454", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:55:01.376780Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 11}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 522, "message_tokens": 429, "message_utf8_bytes": 1176, "cache_tensor_bytes": null, "serialized_payload_bytes": 1176, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11826.122167054564, "ttft_ms": 13.012208044528961, "decode_ms": 132.13691604323685, "receiver_prefill_ms": 64.52366709709167}, "total_wall_ms": 12037.056208122522, "ttft_ms": 13.012208044528961, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5,755,305 CAD", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 0.6666666666666666}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a5a07ccaaa4511595062877d10765d4ada18c53b07b5715cab53137a2f368029", "output_hash": "883728329a58e4233a6d250d9fd306edb2d8c5e65b01ab87194baa37b4105c9c"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "7d00c3195c5ee859a5795337", "attempt_id": "d590d90ebf2c", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:55:01.606522Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 11}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 523, "message_tokens": 429, "message_utf8_bytes": 1176, "cache_tensor_bytes": null, "serialized_payload_bytes": 1176, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11826.122167054564, "ttft_ms": 13.199542067013681, "decode_ms": 13.262416003271937, "receiver_prefill_ms": 64.32179105468094}, "total_wall_ms": 11918.036208022386, "ttft_ms": 13.199542067013681, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a5a07ccaaa4511595062877d10765d4ada18c53b07b5715cab53137a2f368029", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "d81fab21478d8a2773386b9b", "attempt_id": "39311ebeeff3", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:55:01.716945Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 11}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 527, "message_tokens": 429, "message_utf8_bytes": 1176, "cache_tensor_bytes": null, "serialized_payload_bytes": 1176, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11826.122167054564, "ttft_ms": 13.085625017993152, "decode_ms": 119.34249999467283, "receiver_prefill_ms": 63.999458914622664}, "total_wall_ms": 12023.742250050418, "ttft_ms": 13.085625017993152, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5,157,750", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a5a07ccaaa4511595062877d10765d4ada18c53b07b5715cab53137a2f368029", "output_hash": "6324b8c6cb170c034cf4808bbf8ab8cac0afb8556ffd4fe9f871ebb457ba5ffb"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "7ff20243d6ee8f707d03537c", "attempt_id": "d9f9bea0f01a", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:55:01.933103Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 11}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 520, "message_tokens": 429, "message_utf8_bytes": 1176, "cache_tensor_bytes": null, "serialized_payload_bytes": 1176, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11826.122167054564, "ttft_ms": 13.068709056824446, "decode_ms": 118.77633410040289, "receiver_prefill_ms": 63.68250004015863}, "total_wall_ms": 12022.812292096205, "ttft_ms": 13.068709056824446, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5,755,305", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a5a07ccaaa4511595062877d10765d4ada18c53b07b5715cab53137a2f368029", "output_hash": "2feb744bc0efb44064e0bc375d54165760453fdd2c5b159d7c1c555e36e65a43"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "3c14784ce10d599d277815f6", "attempt_id": "272f2b4712d6", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:55:02.147907Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 11}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 529, "message_tokens": 429, "message_utf8_bytes": 1176, "cache_tensor_bytes": null, "serialized_payload_bytes": 1176, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11826.122167054564, "ttft_ms": 13.129958999343216, "decode_ms": 66.2923330673948, "receiver_prefill_ms": 63.87125002220273}, "total_wall_ms": 11970.638417056762, "ttft_ms": 13.129958999343216, "input_usage": {}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Banmunruk-6", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a5a07ccaaa4511595062877d10765d4ada18c53b07b5715cab53137a2f368029", "output_hash": "858197b99462f9923666f025810f174515ef1dac3462c948b1ea68e0be358384"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "5910e33a84b927b5e478e4c7", "attempt_id": "a8093cddf35a", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:55:02.310778Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 11}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 523, "message_tokens": 429, "message_utf8_bytes": 1176, "cache_tensor_bytes": null, "serialized_payload_bytes": 1176, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11826.122167054564, "ttft_ms": 13.18570808507502, "decode_ms": 65.945666981861, "receiver_prefill_ms": 63.446042011491954}, "total_wall_ms": 11969.969042111188, "ttft_ms": 13.18570808507502, "input_usage": {}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "12342", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a5a07ccaaa4511595062877d10765d4ada18c53b07b5715cab53137a2f368029", "output_hash": "307fd56085a6ed891cfb2c38276d7853b122968754378f1153e0127f3a764a0f"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "18c95ccf8f29cbde9d75102f", "attempt_id": "e14f4dbc0021", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:55:02.472895Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 11}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 524, "message_tokens": 429, "message_utf8_bytes": 1176, "cache_tensor_bytes": null, "serialized_payload_bytes": 1176, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11826.122167054564, "ttft_ms": 13.415290974080563, "decode_ms": 13.647375046275556, "receiver_prefill_ms": 64.28379204589874}, "total_wall_ms": 11918.663959018886, "ttft_ms": 13.415290974080563, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a5a07ccaaa4511595062877d10765d4ada18c53b07b5715cab53137a2f368029", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "ced10219f7ad35ad3369a732", "attempt_id": "09d49edc9eec", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:55:02.583531Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 11}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 525, "message_tokens": 429, "message_utf8_bytes": 1176, "cache_tensor_bytes": null, "serialized_payload_bytes": 1176, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11826.122167054564, "ttft_ms": 13.072333997115493, "decode_ms": 13.33312492351979, "receiver_prefill_ms": 64.26291610114276}, "total_wall_ms": 11917.94287506491, "ttft_ms": 13.072333997115493, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a5a07ccaaa4511595062877d10765d4ada18c53b07b5715cab53137a2f368029", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "3ecb7bd2f31cfa91c632e79d", "attempt_id": "ffc72bed164a", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:55:14.523166Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 1356, "message_tokens": 1264, "message_utf8_bytes": 3203, "cache_tensor_bytes": null, "serialized_payload_bytes": 3203, "side_information_bytes": 0, "fetched_bytes": 1849, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11828.981166006997, "ttft_ms": 13.452208018861711, "decode_ms": 93.2539589703083, "receiver_prefill_ms": 144.49370896909386}, "total_wall_ms": 12082.360040978529, "ttft_ms": 13.452208018861711, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5755305", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "26747d9e55400bdf38666d07303298e20117bd64ac16498a8c50e205afa23207", "output_hash": "d62d1e2ffbba79e5e1387defa50c9793b60253cadaf386212b947c0cfb951e11"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "1cc8269a2db3148e4cc0644e", "attempt_id": "d80b3fd52fe5", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:55:14.794611Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 1342, "message_tokens": 1249, "message_utf8_bytes": 3171, "cache_tensor_bytes": null, "serialized_payload_bytes": 3171, "side_information_bytes": 0, "fetched_bytes": 1819, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11828.981166006997, "ttft_ms": 13.262083055451512, "decode_ms": 13.413958949968219, "receiver_prefill_ms": 142.83920894376934}, "total_wall_ms": 12000.646791071631, "ttft_ms": 13.262083055451512, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "6ac0e9e538fbb3f9dd789230af105a8759648afffdbe408253a2634b1d508dd2", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "7a0b8171ac797b3162f5ec07", "attempt_id": "62ce2df48a32", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:55:14.984120Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 1346, "message_tokens": 1249, "message_utf8_bytes": 3171, "cache_tensor_bytes": null, "serialized_payload_bytes": 3171, "side_information_bytes": 0, "fetched_bytes": 1819, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11828.981166006997, "ttft_ms": 13.326582964509726, "decode_ms": 93.65883306600153, "receiver_prefill_ms": 144.37841600738466}, "total_wall_ms": 12082.486958010122, "ttft_ms": 13.326582964509726, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5588115", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "51d492f9476a70b661eedaa1d00b294cf1a79038f3ce9e782b60f428ee01c610", "output_hash": "5e8aa3e317f34247d5545451c63ac9a5b2e481e54d54b520d99a84cd8ee5a8d9"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "f6ab270839d0fcde5aecfc7f", "attempt_id": "6523db16ccff", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:55:15.255559Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 1339, "message_tokens": 1249, "message_utf8_bytes": 3171, "cache_tensor_bytes": null, "serialized_payload_bytes": 3171, "side_information_bytes": 0, "fetched_bytes": 1819, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11828.981166006997, "ttft_ms": 13.809417025186121, "decode_ms": 93.08679203968495, "receiver_prefill_ms": 142.52649992704391}, "total_wall_ms": 12080.547749996185, "ttft_ms": 13.809417025186121, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4439515", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "6ac0e9e538fbb3f9dd789230af105a8759648afffdbe408253a2634b1d508dd2", "output_hash": "428041758b16a119eefbcb06d172852931ea08ef516d1c13ef1c0ae220e36a5a"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "f7d703f48eca3015063d291a", "attempt_id": "70bd86bac3c1", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:55:15.524943Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 1313, "message_tokens": 1214, "message_utf8_bytes": 3140, "cache_tensor_bytes": null, "serialized_payload_bytes": 3140, "side_information_bytes": 0, "fetched_bytes": 1792, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11828.981166006997, "ttft_ms": 13.448332902044058, "decode_ms": 66.47058296948671, "receiver_prefill_ms": 139.71620809752494}, "total_wall_ms": 12050.77024991624, "ttft_ms": 13.448332902044058, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Banmunruk-6", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "6b1b1a48c819fae214c782d2d7b0a8d0bae4addfb962e4000cb4f2725fad17d3", "output_hash": "858197b99462f9923666f025810f174515ef1dac3462c948b1ea68e0be358384"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "110598ac55c9b5d199a03ddc", "attempt_id": "b8ecec5673b0", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:55:15.764840Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 1357, "message_tokens": 1264, "message_utf8_bytes": 3203, "cache_tensor_bytes": null, "serialized_payload_bytes": 3203, "side_information_bytes": 0, "fetched_bytes": 1849, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11828.981166006997, "ttft_ms": 13.481791946105659, "decode_ms": 106.81537503842264, "receiver_prefill_ms": 144.02137498836964}, "total_wall_ms": 12095.443791011348, "ttft_ms": 13.481791946105659, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "18818407", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "26747d9e55400bdf38666d07303298e20117bd64ac16498a8c50e205afa23207", "output_hash": "55560e5389b87684bbf836f2721aa4a83ea33c3e9bbbf6fdc232d88976a40349"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "1679b46af22f1914954ed320", "attempt_id": "d7345499e065", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:55:16.049266Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 1343, "message_tokens": 1249, "message_utf8_bytes": 3171, "cache_tensor_bytes": null, "serialized_payload_bytes": 3171, "side_information_bytes": 0, "fetched_bytes": 1819, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11828.981166006997, "ttft_ms": 13.310667010955513, "decode_ms": 13.248000061139464, "receiver_prefill_ms": 143.03866703994572}, "total_wall_ms": 12000.677624018863, "ttft_ms": 13.310667010955513, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "6ac0e9e538fbb3f9dd789230af105a8759648afffdbe408253a2634b1d508dd2", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "bc476538f2d14d4c9cbf5e8d", "attempt_id": "23ac5b0944d0", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:55:16.238709Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 1344, "message_tokens": 1249, "message_utf8_bytes": 3171, "cache_tensor_bytes": null, "serialized_payload_bytes": 3171, "side_information_bytes": 0, "fetched_bytes": 1819, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11828.981166006997, "ttft_ms": 13.670457992702723, "decode_ms": 13.240958098322153, "receiver_prefill_ms": 142.73533399682492}, "total_wall_ms": 12000.745832920074, "ttft_ms": 13.670457992702723, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "6ac0e9e538fbb3f9dd789230af105a8759648afffdbe408253a2634b1d508dd2", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "72472828a0d109975db8e9ff", "attempt_id": "d080c8d0af7f", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:55:17.116857Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 228151296}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 46, "message_tokens": null, "message_utf8_bytes": 30180, "cache_tensor_bytes": 228151296, "serialized_payload_bytes": 228151296, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 688.2733750389889, "ttft_ms": 48.594582942314446, "decode_ms": 105.02450005151331, "receiver_prefill_ms": 0.0}, "total_wall_ms": 851.1960420291871, "ttft_ms": 48.594582942314446, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5755305", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "d62d1e2ffbba79e5e1387defa50c9793b60253cadaf386212b947c0cfb951e11"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "472070c53de99f00ce749975", "attempt_id": "aa105f5fbf13", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:55:17.297478Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 228151296}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 30180, "cache_tensor_bytes": 228151296, "serialized_payload_bytes": 228151296, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 688.2733750389889, "ttft_ms": 45.20816693548113, "decode_ms": 149.39775003585964, "receiver_prefill_ms": 0.0}, "total_wall_ms": 890.2958331163973, "ttft_ms": 45.20816693548113, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "GY-13E-6844", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "2ca6ddf54be9e197af2f12973eddba19b3d0c78d0073a82218a23e1d9bbad92e"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "606a483c6e1bb50db6d3635a", "attempt_id": "c0ff618b5bee", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:55:17.517474Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 228151296}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 51, "message_tokens": null, "message_utf8_bytes": 30180, "cache_tensor_bytes": 228151296, "serialized_payload_bytes": 228151296, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 688.2733750389889, "ttft_ms": 44.15058298036456, "decode_ms": 134.61162499152124, "receiver_prefill_ms": 0.0}, "total_wall_ms": 873.6185000743717, "ttft_ms": 44.15058298036456, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "6,528,610", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "f6c6398ddb4eb11af59cf7b1b2092123f645445d9e3810c1fd3abca4cab07946"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "91871147efb638da37d95aed", "attempt_id": "82fd7e587c3e", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:55:17.720931Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 228151296}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 44, "message_tokens": null, "message_utf8_bytes": 30180, "cache_tensor_bytes": 228151296, "serialized_payload_bytes": 228151296, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 688.2733750389889, "ttft_ms": 47.48945799656212, "decode_ms": 134.74087498616427, "receiver_prefill_ms": 0.0}, "total_wall_ms": 878.2309159869328, "ttft_ms": 47.48945799656212, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5,968,405", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "2a7eb75269096d6b1f6cf3dd00d83539da50b512010957b3898ea486c882f19c"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "0c91381cb2c2c946a8630608", "attempt_id": "5c1d99f5d66c", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:55:17.928820Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 228151296}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 53, "message_tokens": null, "message_utf8_bytes": 30180, "cache_tensor_bytes": 228151296, "serialized_payload_bytes": 228151296, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 688.2733750389889, "ttft_ms": 47.189958044327796, "decode_ms": 75.28345799073577, "receiver_prefill_ms": 0.0}, "total_wall_ms": 818.4112501330674, "ttft_ms": 47.189958044327796, "input_usage": {}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Banmunruk-6", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "858197b99462f9923666f025810f174515ef1dac3462c948b1ea68e0be358384"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "e3d304857de69faf3062912b", "attempt_id": "2ee33dea09d5", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:55:18.076827Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 228151296}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 30180, "cache_tensor_bytes": 228151296, "serialized_payload_bytes": 228151296, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 688.2733750389889, "ttft_ms": 45.55049992632121, "decode_ms": 119.05941704753786, "receiver_prefill_ms": 0.0}, "total_wall_ms": 860.4637919925153, "ttft_ms": 45.55049992632121, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "17414416", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "172b87f2118ddfeb841852240360e4d1639e0383bb85a5865cc5819afcd89eef"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a976267a023f6bcf3ed869c8", "attempt_id": "4814c7dee597", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:55:18.267710Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 228151296}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 48, "message_tokens": null, "message_utf8_bytes": 30180, "cache_tensor_bytes": 228151296, "serialized_payload_bytes": 228151296, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 688.2733750389889, "ttft_ms": 48.29437506850809, "decode_ms": 14.833666966296732, "receiver_prefill_ms": 0.0}, "total_wall_ms": 758.8569161016494, "ttft_ms": 48.29437506850809, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "88aa8e11ee9bdab2766aa387", "attempt_id": "feb3ea85543e", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:55:18.356528Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 228151296}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 49, "message_tokens": null, "message_utf8_bytes": 30180, "cache_tensor_bytes": 228151296, "serialized_payload_bytes": 228151296, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 688.2733750389889, "ttft_ms": 44.902749941684306, "decode_ms": 15.005249995738268, "receiver_prefill_ms": 0.0}, "total_wall_ms": 755.535583011806, "ttft_ms": 44.902749941684306, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "daa7dccf85ca7858d100420c", "attempt_id": "0eb10fe3b495", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:55:18.442116Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 6313, "message_tokens": 6225, "message_utf8_bytes": 30710, "cache_tensor_bytes": null, "serialized_payload_bytes": 30710, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.04600000102072954, "ttft_ms": 13.97208299022168, "decode_ms": 95.77641601208597, "receiver_prefill_ms": 740.4797499766573}, "total_wall_ms": 859.7526659723371, "ttft_ms": 13.97208299022168, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5755305", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "e84780918aa493b32fe8fdc658dcacc801e4dd8b3c91ff114bb07989574c1d9a", "output_hash": "d62d1e2ffbba79e5e1387defa50c9793b60253cadaf386212b947c0cfb951e11"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "8ae3c455407843506e546a82", "attempt_id": "4b3d8f043a9e", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:55:19.319478Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 6314, "message_tokens": 6225, "message_utf8_bytes": 30710, "cache_tensor_bytes": null, "serialized_payload_bytes": 30710, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.04600000102072954, "ttft_ms": 13.653958914801478, "decode_ms": 136.3406670279801, "receiver_prefill_ms": 743.477875017561}, "total_wall_ms": 903.2056250143796, "ttft_ms": 13.653958914801478, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "GY-13E-6844", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "e84780918aa493b32fe8fdc658dcacc801e4dd8b3c91ff114bb07989574c1d9a", "output_hash": "2ca6ddf54be9e197af2f12973eddba19b3d0c78d0073a82218a23e1d9bbad92e"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "04acd6d057292e5cdebe7a0d", "attempt_id": "483a61bdb4ae", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:55:20.240342Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 6318, "message_tokens": 6225, "message_utf8_bytes": 30710, "cache_tensor_bytes": null, "serialized_payload_bytes": 30710, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.04600000102072954, "ttft_ms": 13.735625077970326, "decode_ms": 95.88695806451142, "receiver_prefill_ms": 745.3571250662208}, "total_wall_ms": 864.5172500982881, "ttft_ms": 13.735625077970326, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5588115", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "e84780918aa493b32fe8fdc658dcacc801e4dd8b3c91ff114bb07989574c1d9a", "output_hash": "5e8aa3e317f34247d5545451c63ac9a5b2e481e54d54b520d99a84cd8ee5a8d9"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "d00bd17eed6d22034dffbac9", "attempt_id": "802af20bcf7d", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:55:21.122978Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 6311, "message_tokens": 6225, "message_utf8_bytes": 30710, "cache_tensor_bytes": null, "serialized_payload_bytes": 30710, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.04600000102072954, "ttft_ms": 13.847583089955151, "decode_ms": 96.35524998884648, "receiver_prefill_ms": 750.235041952692}, "total_wall_ms": 870.086667011492, "ttft_ms": 13.847583089955151, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4439515", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "e84780918aa493b32fe8fdc658dcacc801e4dd8b3c91ff114bb07989574c1d9a", "output_hash": "428041758b16a119eefbcb06d172852931ea08ef516d1c13ef1c0ae220e36a5a"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "6ca5cda8d715ea4118872fb8", "attempt_id": "59af01bba3a3", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:55:22.010828Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 6320, "message_tokens": 6225, "message_utf8_bytes": 30710, "cache_tensor_bytes": null, "serialized_payload_bytes": 30710, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.04600000102072954, "ttft_ms": 13.890875037759542, "decode_ms": 68.07491602376103, "receiver_prefill_ms": 755.2366249728948}, "total_wall_ms": 846.8017920386046, "ttft_ms": 13.890875037759542, "input_usage": {}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Banmunruk-6", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "e84780918aa493b32fe8fdc658dcacc801e4dd8b3c91ff114bb07989574c1d9a", "output_hash": "858197b99462f9923666f025810f174515ef1dac3462c948b1ea68e0be358384"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "f9a3be3f4c120b3a8851ab51", "attempt_id": "4376fcd8911b", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:55:22.875089Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 6314, "message_tokens": 6225, "message_utf8_bytes": 30710, "cache_tensor_bytes": null, "serialized_payload_bytes": 30710, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.04600000102072954, "ttft_ms": 14.07337503042072, "decode_ms": 110.62250006943941, "receiver_prefill_ms": 760.6123329605907}, "total_wall_ms": 894.9564999202266, "ttft_ms": 14.07337503042072, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "21408819", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "e84780918aa493b32fe8fdc658dcacc801e4dd8b3c91ff114bb07989574c1d9a", "output_hash": "81559d35522cf7fccef9a8b0321cd2e4fc4e6946933dc2aacd8eed38f0c9613d"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "88da563100cdfb375107799f", "attempt_id": "58dead1add8a", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:55:23.788134Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 6315, "message_tokens": 6225, "message_utf8_bytes": 30710, "cache_tensor_bytes": null, "serialized_payload_bytes": 30710, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.04600000102072954, "ttft_ms": 13.8547089882195, "decode_ms": 14.112750068306923, "receiver_prefill_ms": 769.8565419996157}, "total_wall_ms": 807.5896670343354, "ttft_ms": 13.8547089882195, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "e84780918aa493b32fe8fdc658dcacc801e4dd8b3c91ff114bb07989574c1d9a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "d1028affdef9d70e8de53988", "attempt_id": "a97e8e0eb84f", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:55:24.613341Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 6316, "message_tokens": 6225, "message_utf8_bytes": 30710, "cache_tensor_bytes": null, "serialized_payload_bytes": 30710, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.04600000102072954, "ttft_ms": 14.324249932542443, "decode_ms": 13.854249962605536, "receiver_prefill_ms": 795.693084015511}, "total_wall_ms": 833.9814170030877, "ttft_ms": 14.324249932542443, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "e84780918aa493b32fe8fdc658dcacc801e4dd8b3c91ff114bb07989574c1d9a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c206efce1850f1025e94f286", "attempt_id": "b2798d7bcbbb", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:55:51.257907Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1876}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 1991, "message_tokens": 1901, "message_utf8_bytes": 6825, "cache_tensor_bytes": null, "serialized_payload_bytes": 6825, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25792.245708056726, "ttft_ms": 13.500417000614107, "decode_ms": 94.19654100202024, "receiver_prefill_ms": 209.19295796193182}, "total_wall_ms": 26112.448667059653, "ttft_ms": 13.500417000614107, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5755305", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ddeab8f4241d7fe74f98340601b6c66e3916d7fa1005e7822eb8162d283819f7", "output_hash": "d62d1e2ffbba79e5e1387defa50c9793b60253cadaf386212b947c0cfb951e11"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "722e8b6405ba96158868080e", "attempt_id": "44d639d9ab53", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:55:51.596876Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1876}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 1992, "message_tokens": 1901, "message_utf8_bytes": 6825, "cache_tensor_bytes": null, "serialized_payload_bytes": 6825, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25792.245708056726, "ttft_ms": 13.840208994224668, "decode_ms": 136.7079159244895, "receiver_prefill_ms": 230.03433307167143}, "total_wall_ms": 26176.05204100255, "ttft_ms": 13.840208994224668, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "GY-13E-6844", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ddeab8f4241d7fe74f98340601b6c66e3916d7fa1005e7822eb8162d283819f7", "output_hash": "2ca6ddf54be9e197af2f12973eddba19b3d0c78d0073a82218a23e1d9bbad92e"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "f370e779b5fe6462a32f9520", "attempt_id": "7305d0858645", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:55:51.999139Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1876}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 1996, "message_tokens": 1901, "message_utf8_bytes": 6825, "cache_tensor_bytes": null, "serialized_payload_bytes": 6825, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25792.245708056726, "ttft_ms": 13.828916940838099, "decode_ms": 95.0315409572795, "receiver_prefill_ms": 268.28879199456424}, "total_wall_ms": 26172.586083062924, "ttft_ms": 13.828916940838099, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5588115", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ddeab8f4241d7fe74f98340601b6c66e3916d7fa1005e7822eb8162d283819f7", "output_hash": "5e8aa3e317f34247d5545451c63ac9a5b2e481e54d54b520d99a84cd8ee5a8d9"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "7554ff2d3ec9b813d339cc69", "attempt_id": "2be32229fefe", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:55:52.397832Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1876}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 1989, "message_tokens": 1901, "message_utf8_bytes": 6825, "cache_tensor_bytes": null, "serialized_payload_bytes": 6825, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25792.245708056726, "ttft_ms": 13.845833949744701, "decode_ms": 123.16758290398866, "receiver_prefill_ms": 267.58608408272266}, "total_wall_ms": 26200.139333028346, "ttft_ms": 13.845833949744701, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4,439,515", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ddeab8f4241d7fe74f98340601b6c66e3916d7fa1005e7822eb8162d283819f7", "output_hash": "7e209a4c9dd6879d7024b31aa971d69ea31077909c7ea20bb771181ec5b64af9"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "7d2ac955d0fb4ba94273f212", "attempt_id": "f4a64fc7a871", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:55:52.824393Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1876}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 1998, "message_tokens": 1901, "message_utf8_bytes": 6825, "cache_tensor_bytes": null, "serialized_payload_bytes": 6825, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25792.245708056726, "ttft_ms": 14.086958020925522, "decode_ms": 68.51608399301767, "receiver_prefill_ms": 255.63466700259596}, "total_wall_ms": 26133.777083014138, "ttft_ms": 14.086958020925522, "input_usage": {}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Banmunruk-6", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ddeab8f4241d7fe74f98340601b6c66e3916d7fa1005e7822eb8162d283819f7", "output_hash": "858197b99462f9923666f025810f174515ef1dac3462c948b1ea68e0be358384"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "404171aa03b4af8e95e6aa66", "attempt_id": "d92dde420b20", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:55:53.184367Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1876}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 1992, "message_tokens": 1901, "message_utf8_bytes": 6825, "cache_tensor_bytes": null, "serialized_payload_bytes": 6825, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25792.245708056726, "ttft_ms": 14.100999920628965, "decode_ms": 136.08841598033905, "receiver_prefill_ms": 257.21187493763864}, "total_wall_ms": 26202.809833106585, "ttft_ms": 14.100999920628965, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "20,497,256", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ddeab8f4241d7fe74f98340601b6c66e3916d7fa1005e7822eb8162d283819f7", "output_hash": "98e8d6c180c4690fa5e533a1d5056d233b372603417a0de92a742e1c66604f96"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "1f1ef95ed62c1096315c8b64", "attempt_id": "c7181d87b159", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:55:53.614013Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1876}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 1993, "message_tokens": 1901, "message_utf8_bytes": 6825, "cache_tensor_bytes": null, "serialized_payload_bytes": 6825, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25792.245708056726, "ttft_ms": 13.614541036076844, "decode_ms": 13.578874990344048, "receiver_prefill_ms": 241.4399170083925}, "total_wall_ms": 26063.966374029405, "ttft_ms": 13.614541036076844, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ddeab8f4241d7fe74f98340601b6c66e3916d7fa1005e7822eb8162d283819f7", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "5d5c7bce879f113c570b729c", "attempt_id": "a5cd26d5cbe2", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:55:53.903821Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1876}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 1994, "message_tokens": 1901, "message_utf8_bytes": 6825, "cache_tensor_bytes": null, "serialized_payload_bytes": 6825, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25792.245708056726, "ttft_ms": 13.788875075988472, "decode_ms": 13.932124944403768, "receiver_prefill_ms": 255.39199996273965}, "total_wall_ms": 26078.57470808085, "ttft_ms": 13.788875075988472, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ddeab8f4241d7fe74f98340601b6c66e3916d7fa1005e7822eb8162d283819f7", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "31037387960393c45c9d6b3a", "attempt_id": "937ebe9bfadb", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:56:20.516057Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 27}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 1147, "message_tokens": 1054, "message_utf8_bytes": 2805, "cache_tensor_bytes": null, "serialized_payload_bytes": 2805, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26307.385249994695, "ttft_ms": 19.89424997009337, "decode_ms": 14.355999999679625, "receiver_prefill_ms": 1320.0275410199538}, "total_wall_ms": 27663.6082920013, "ttft_ms": 19.89424997009337, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "4dd9f525c23ad3f679f9df6ed77e7a25e553831a3a9c7785862ce8f303eb15d9", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "7e72e4d2ee67289a025f52a9", "attempt_id": "e90862367db8", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:56:21.900583Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 27}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 1148, "message_tokens": 1054, "message_utf8_bytes": 2805, "cache_tensor_bytes": null, "serialized_payload_bytes": 2805, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26307.385249994695, "ttft_ms": 13.927667052485049, "decode_ms": 13.630999950692058, "receiver_prefill_ms": 425.65462505444884}, "total_wall_ms": 26762.714959098957, "ttft_ms": 13.927667052485049, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "4dd9f525c23ad3f679f9df6ed77e7a25e553831a3a9c7785862ce8f303eb15d9", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b790d0adf31ca47e167b843e", "attempt_id": "7eca60e090e4", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:56:22.376187Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 27}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 1152, "message_tokens": 1054, "message_utf8_bytes": 2805, "cache_tensor_bytes": null, "serialized_payload_bytes": 2805, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26307.385249994695, "ttft_ms": 13.838291983120143, "decode_ms": 97.31583308894187, "receiver_prefill_ms": 132.8432500595227}, "total_wall_ms": 26553.434500005096, "ttft_ms": 13.838291983120143, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5588115", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "4dd9f525c23ad3f679f9df6ed77e7a25e553831a3a9c7785862ce8f303eb15d9", "output_hash": "5e8aa3e317f34247d5545451c63ac9a5b2e481e54d54b520d99a84cd8ee5a8d9"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c70620689fde14e8ee1a06e4", "attempt_id": "6c89163341b4", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:56:22.649770Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 27}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 1145, "message_tokens": 1054, "message_utf8_bytes": 2805, "cache_tensor_bytes": null, "serialized_payload_bytes": 2805, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26307.385249994695, "ttft_ms": 14.012082945555449, "decode_ms": 124.88254194613546, "receiver_prefill_ms": 121.8432089081034}, "total_wall_ms": 26570.102124940604, "ttft_ms": 14.012082945555449, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4,926,635", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "4dd9f525c23ad3f679f9df6ed77e7a25e553831a3a9c7785862ce8f303eb15d9", "output_hash": "d5d47d521beaf07dbde6875769c832c6ec5c9a76e4045224352cc0abc293f5d5"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "aa21c223d9ff9ce84e6609e3", "attempt_id": "c1c1489127e9", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:56:22.931398Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 27}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 1154, "message_tokens": 1054, "message_utf8_bytes": 2805, "cache_tensor_bytes": null, "serialized_payload_bytes": 2805, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26307.385249994695, "ttft_ms": 13.517750077880919, "decode_ms": 67.53641704563051, "receiver_prefill_ms": 127.04112497158349}, "total_wall_ms": 26517.318958067335, "ttft_ms": 13.517750077880919, "input_usage": {}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Banmunruk-6", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "4dd9f525c23ad3f679f9df6ed77e7a25e553831a3a9c7785862ce8f303eb15d9", "output_hash": "858197b99462f9923666f025810f174515ef1dac3462c948b1ea68e0be358384"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "7b4568a97338df3e5525bf08", "attempt_id": "01868c4fddcf", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:56:23.159208Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 27}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 1148, "message_tokens": 1054, "message_utf8_bytes": 2805, "cache_tensor_bytes": null, "serialized_payload_bytes": 2805, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26307.385249994695, "ttft_ms": 13.52216606028378, "decode_ms": 93.73808302916586, "receiver_prefill_ms": 136.3510829396546}, "total_wall_ms": 26552.710457937792, "ttft_ms": 13.52216606028378, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1140705", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "4dd9f525c23ad3f679f9df6ed77e7a25e553831a3a9c7785862ce8f303eb15d9", "output_hash": "6b3c3572ebfa74b242333c32f0e13eb8c15b7941d6cd84768bb67cff7fe1079d"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "415e808883a82d726f3c7ec5", "attempt_id": "7ca9167a8ab1", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:56:23.422854Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 27}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 1149, "message_tokens": 1054, "message_utf8_bytes": 2805, "cache_tensor_bytes": null, "serialized_payload_bytes": 2805, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26307.385249994695, "ttft_ms": 13.58866598457098, "decode_ms": 13.530625030398369, "receiver_prefill_ms": 139.40604194067419}, "total_wall_ms": 26475.710541009903, "ttft_ms": 13.58866598457098, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "4dd9f525c23ad3f679f9df6ed77e7a25e553831a3a9c7785862ce8f303eb15d9", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "96e39b70e36aebd2d800344f", "attempt_id": "5daf8b9b3b2f", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:56:23.609476Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 27}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 1150, "message_tokens": 1054, "message_utf8_bytes": 2805, "cache_tensor_bytes": null, "serialized_payload_bytes": 2805, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26307.385249994695, "ttft_ms": 13.720959075726569, "decode_ms": 13.892042101360857, "receiver_prefill_ms": 193.2805839460343}, "total_wall_ms": 26530.029791989364, "ttft_ms": 13.720959075726569, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "4dd9f525c23ad3f679f9df6ed77e7a25e553831a3a9c7785862ce8f303eb15d9", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "df0b9c738a15275621ed6920", "attempt_id": "9544d0680319", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:56:50.268893Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 2683, "message_tokens": 2591, "message_utf8_bytes": 6721, "cache_tensor_bytes": null, "serialized_payload_bytes": 6721, "side_information_bytes": 0, "fetched_bytes": 3541, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26419.22387504019, "ttft_ms": 14.067583018913865, "decode_ms": 97.18008292838931, "receiver_prefill_ms": 293.80554193630815}, "total_wall_ms": 26829.39716707915, "ttft_ms": 14.067583018913865, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5755305", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "9fc533cdeb0646df39bf883dbcc9ee1810177f99f32126bb80af5d01c267acda", "output_hash": "d62d1e2ffbba79e5e1387defa50c9793b60253cadaf386212b947c0cfb951e11"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "440b6a3ba2c8f1590934e445", "attempt_id": "ce07e28a14e2", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:56:50.705431Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 2484, "message_tokens": 2391, "message_utf8_bytes": 6376, "cache_tensor_bytes": null, "serialized_payload_bytes": 6376, "side_information_bytes": 0, "fetched_bytes": 3217, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26419.22387504019, "ttft_ms": 14.709499897435308, "decode_ms": 138.83283303584903, "receiver_prefill_ms": 385.16983296722174}, "total_wall_ms": 26962.200416019186, "ttft_ms": 14.709499897435308, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "FN-61P-4981", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "5aa850d3be083a946b296f739484e7326694b44831b2580da2656502fc980b5f", "output_hash": "13e8328779a360703521a138305d0c3e1a42a5adbece3e3c52042c4a3ea91de2"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "e840992739caa9a7e3477d17", "attempt_id": "6b4e6d138097", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:56:51.271279Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 2685, "message_tokens": 2588, "message_utf8_bytes": 6644, "cache_tensor_bytes": null, "serialized_payload_bytes": 6644, "side_information_bytes": 0, "fetched_bytes": 3474, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26419.22387504019, "ttft_ms": 14.411707990802824, "decode_ms": 96.27062489744276, "receiver_prefill_ms": 360.253082937561}, "total_wall_ms": 26893.59841600526, "ttft_ms": 14.411707990802824, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5588115", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "7a928bf9b5ab2e66fd21b2e947080ea60db6ccdd118b6d53a998a1acddb76dc0", "output_hash": "5e8aa3e317f34247d5545451c63ac9a5b2e481e54d54b520d99a84cd8ee5a8d9"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "99a37c3c2d2cf0a1cd0f83b3", "attempt_id": "06e72171cedb", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:56:51.764053Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 2678, "message_tokens": 2588, "message_utf8_bytes": 6644, "cache_tensor_bytes": null, "serialized_payload_bytes": 6644, "side_information_bytes": 0, "fetched_bytes": 3474, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26419.22387504019, "ttft_ms": 14.363750000484288, "decode_ms": 96.66920895688236, "receiver_prefill_ms": 352.42612508591264}, "total_wall_ms": 26886.13100000657, "ttft_ms": 14.363750000484288, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4439515", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "7a928bf9b5ab2e66fd21b2e947080ea60db6ccdd118b6d53a998a1acddb76dc0", "output_hash": "428041758b16a119eefbcb06d172852931ea08ef516d1c13ef1c0ae220e36a5a"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "d83695f9445b9e9ce46fe57d", "attempt_id": "82c358993843", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:56:52.248618Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 2566, "message_tokens": 2467, "message_utf8_bytes": 6451, "cache_tensor_bytes": null, "serialized_payload_bytes": 6451, "side_information_bytes": 0, "fetched_bytes": 3306, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26419.22387504019, "ttft_ms": 13.463874929584563, "decode_ms": 67.63395911548287, "receiver_prefill_ms": 608.9850830612704}, "total_wall_ms": 27113.419624976814, "ttft_ms": 13.463874929584563, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Banmunruk-6", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "13a4c0273734934277c99b6eee3e09077f18c9f6c02bd280355247bc53db36a9", "output_hash": "858197b99462f9923666f025810f174515ef1dac3462c948b1ea68e0be358384"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "e76d4bf2b56429ee608caac2", "attempt_id": "cfd0651cfbdd", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:56:52.961406Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 2676, "message_tokens": 2583, "message_utf8_bytes": 6678, "cache_tensor_bytes": null, "serialized_payload_bytes": 6678, "side_information_bytes": 0, "fetched_bytes": 3503, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26419.22387504019, "ttft_ms": 14.088292024098337, "decode_ms": 95.56620800867677, "receiver_prefill_ms": 351.77341604139656}, "total_wall_ms": 26884.195624967106, "ttft_ms": 14.088292024098337, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1748405", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "c856097e223617f302df46521a33cf6b74c444a56a9c870cb5526d869371a809", "output_hash": "7e279b2b8c3695908a1c2ec9912b479a325b8179b6d974b0610133f9d4cbf681"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "d68629117e009308cba26efa", "attempt_id": "bbaeb6d8c19b", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:56:53.444435Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 2599, "message_tokens": 2505, "message_utf8_bytes": 6621, "cache_tensor_bytes": null, "serialized_payload_bytes": 6621, "side_information_bytes": 0, "fetched_bytes": 3422, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26419.22387504019, "ttft_ms": 13.766542077064514, "decode_ms": 13.607083004899323, "receiver_prefill_ms": 302.73012490943074}, "total_wall_ms": 26752.745417063124, "ttft_ms": 13.766542077064514, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a1915cc4ebd323437c62c297168a90b90edd2fbe9da400b694341881a5db72f9", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b422f91884bc497c996ebc21", "attempt_id": "4674aff46af1", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:56:53.795926Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 2691, "message_tokens": 2596, "message_utf8_bytes": 6687, "cache_tensor_bytes": null, "serialized_payload_bytes": 6687, "side_information_bytes": 0, "fetched_bytes": 3512, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26419.22387504019, "ttft_ms": 14.200582983903587, "decode_ms": 13.958041905425489, "receiver_prefill_ms": 389.61320801172405}, "total_wall_ms": 26840.650791069493, "ttft_ms": 14.200582983903587, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "4d371e068ad5c2a3973ec9ae4f71cae43810b22cb28295a3134fd8ced6f74bac", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "e11a7210c2857a32f9ffbdbd", "attempt_id": "0679171e2bc7", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:56:55.923918Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 235229184}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 46, "message_tokens": null, "message_utf8_bytes": 30656, "cache_tensor_bytes": 235229184, "serialized_payload_bytes": 235229184, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 1688.6042089900002, "ttft_ms": 65.86462503764778, "decode_ms": 141.2853329675272, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1958.545958972536, "ttft_ms": 65.86462503764778, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5,755,305", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "2feb744bc0efb44064e0bc375d54165760453fdd2c5b159d7c1c555e36e65a43"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "bbc436082f03a6de1434dfde", "attempt_id": "a5ff5d63b8ee", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:56:56.212845Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 235229184}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 30656, "cache_tensor_bytes": 235229184, "serialized_payload_bytes": 235229184, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 1688.6042089900002, "ttft_ms": 96.11458401195705, "decode_ms": 149.82191601302475, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1942.5389169482514, "ttft_ms": 96.11458401195705, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "GY-13E-6844", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "2ca6ddf54be9e197af2f12973eddba19b3d0c78d0073a82218a23e1d9bbad92e"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "aa610b35b9033566e69263aa", "attempt_id": "247d9a7257ee", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:56:56.489738Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 235229184}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 51, "message_tokens": null, "message_utf8_bytes": 30656, "cache_tensor_bytes": 235229184, "serialized_payload_bytes": 235229184, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 1688.6042089900002, "ttft_ms": 84.44858307484537, "decode_ms": 134.49349999427795, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1911.222376045771, "ttft_ms": 84.44858307484537, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5,588,115", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "6e5b387355d2edb1ffd589b58491594fa792b2b6e99363b2e5710b42eb6429b0"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a4df1df19112df82ad152768", "attempt_id": "3c06e3177edc", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:56:56.730275Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 235229184}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 44, "message_tokens": null, "message_utf8_bytes": 30656, "cache_tensor_bytes": 235229184, "serialized_payload_bytes": 235229184, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 1688.6042089900002, "ttft_ms": 100.01279204152524, "decode_ms": 135.31345897354186, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1928.4377919975668, "ttft_ms": 100.01279204152524, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5,968,405", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "2a7eb75269096d6b1f6cf3dd00d83539da50b512010957b3898ea486c882f19c"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "ec8f84d5fb5ab3e70e0a06dc", "attempt_id": "eae1ddc7514f", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:56:56.987706Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 235229184}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 53, "message_tokens": null, "message_utf8_bytes": 30656, "cache_tensor_bytes": 235229184, "serialized_payload_bytes": 235229184, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 1688.6042089900002, "ttft_ms": 83.67137494497001, "decode_ms": 75.6885000737384, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1851.6985840396956, "ttft_ms": 83.67137494497001, "input_usage": {}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Banmunruk-6", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "858197b99462f9923666f025810f174515ef1dac3462c948b1ea68e0be358384"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "132040a58ede3f1ea50d146b", "attempt_id": "a0245ed54255", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:56:57.168485Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 235229184}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 30656, "cache_tensor_bytes": 235229184, "serialized_payload_bytes": 235229184, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 1688.6042089900002, "ttft_ms": 46.44887498579919, "decode_ms": 119.63491700589657, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1859.2416669707745, "ttft_ms": 46.44887498579919, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "12304917", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "7f9a80add55fd124f17ce8e16e9f2a536000b06df6554b03569a65be20086fa7"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "79f5e967918e765c1b236474", "attempt_id": "2d99e372ea3f", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:56:57.356864Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 235229184}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 48, "message_tokens": null, "message_utf8_bytes": 30656, "cache_tensor_bytes": 235229184, "serialized_payload_bytes": 235229184, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 1688.6042089900002, "ttft_ms": 49.28687494248152, "decode_ms": 15.308916918002069, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1757.5214999960735, "ttft_ms": 49.28687494248152, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "ea30159222a3b0294e5f7f6a", "attempt_id": "92377d0c02ca", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:56:57.443672Z", "task_family": "synthetic", "document_id": "syn_03001_8000", "document_hash": "bada8ba4956b59f8a22a68843c4c56f104029ebc33a4c35fcf192e0cf7667914", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 235229184}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5789, "query_tokens": 49, "message_tokens": null, "message_utf8_bytes": 30656, "cache_tensor_bytes": 235229184, "serialized_payload_bytes": 235229184, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 1688.6042089900002, "ttft_ms": 46.244541998021305, "decode_ms": 15.027291956357658, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1754.4155839132145, "ttft_ms": 46.244541998021305, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Briselmun-8", "bogus_amount": "5,550,615", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03001_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "fa323b79a4aac360f9abfce6", "attempt_id": "60a2a0bc84f1", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:56:58.508246Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 2263, "message_tokens": 2177, "message_utf8_bytes": 8286, "cache_tensor_bytes": null, "serialized_payload_bytes": 8286, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.033250078558921814, "ttft_ms": 13.679625000804663, "decode_ms": 93.41291594319046, "receiver_prefill_ms": 260.0462919799611}, "total_wall_ms": 371.0092920809984, "ttft_ms": 13.679625000804663, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5863814", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "1acde7a06be99d32a938c209eb287ec81faba3dedb13cc43c3a7e23397c698ee", "output_hash": "8a4efc21da7fd0aafdbb1f8f5fe9231f746cd4dd77afd1da76a2f1ed0b271b75"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "01bc7c432fc73bcc162fb8e2", "attempt_id": "39f857a0072b", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:56:58.895095Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 2265, "message_tokens": 2177, "message_utf8_bytes": 8286, "cache_tensor_bytes": null, "serialized_payload_bytes": 8286, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.033250078558921814, "ttft_ms": 13.396249967627227, "decode_ms": 134.2596251051873, "receiver_prefill_ms": 248.16508300136775}, "total_wall_ms": 399.6490001445636, "ttft_ms": 13.396249967627227, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "NK-95Z-2165", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "1acde7a06be99d32a938c209eb287ec81faba3dedb13cc43c3a7e23397c698ee", "output_hash": "bd8f809b02304c6b9dc7c1e116ecccc2328c976a5c071e89a94d661b48e73aa7"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a8cc4ed7b7226f37b7e68973", "attempt_id": "6fea71fc62ea", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:56:59.313809Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 2272, "message_tokens": 2177, "message_utf8_bytes": 8286, "cache_tensor_bytes": null, "serialized_payload_bytes": 8286, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.033250078558921814, "ttft_ms": 13.416499947197735, "decode_ms": 93.09220896102488, "receiver_prefill_ms": 248.3887920388952}, "total_wall_ms": 358.39120810851455, "ttft_ms": 13.416499947197735, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "848,090", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "1acde7a06be99d32a938c209eb287ec81faba3dedb13cc43c3a7e23397c698ee", "output_hash": "b5105383225daa63c1554bd08d37339a9b2d2a1066058fc15ab16cf75f05d898"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "342c201619915f5cfe3c58a1", "attempt_id": "7f85e98470d8", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:56:59.686613Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 2265, "message_tokens": 2177, "message_utf8_bytes": 8286, "cache_tensor_bytes": null, "serialized_payload_bytes": 8286, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.033250078558921814, "ttft_ms": 13.874541968107224, "decode_ms": 95.50329204648733, "receiver_prefill_ms": 248.24795895256102}, "total_wall_ms": 361.08120810240507, "ttft_ms": 13.874541968107224, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "848,090", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "1acde7a06be99d32a938c209eb287ec81faba3dedb13cc43c3a7e23397c698ee", "output_hash": "b5105383225daa63c1554bd08d37339a9b2d2a1066058fc15ab16cf75f05d898"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "604eeb77785c8b5a4fdf5c11", "attempt_id": "ac4e25c7df87", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:57:00.061552Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 2272, "message_tokens": 2177, "message_utf8_bytes": 8286, "cache_tensor_bytes": null, "serialized_payload_bytes": 8286, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.033250078558921814, "ttft_ms": 13.477375032380223, "decode_ms": 80.06862504407763, "receiver_prefill_ms": 248.23658308014274}, "total_wall_ms": 345.4081250820309, "ttft_ms": 13.477375032380223, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Gamtiszen-2", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "1acde7a06be99d32a938c209eb287ec81faba3dedb13cc43c3a7e23397c698ee", "output_hash": "32e3af428886434823d82102e1c482765eafdf6077fe8d3fba5752ec5e29826c"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "16c5483cb2d2a1145813b069", "attempt_id": "f124bcd705c3", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:57:00.424292Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 2266, "message_tokens": 2177, "message_utf8_bytes": 8286, "cache_tensor_bytes": null, "serialized_payload_bytes": 8286, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.033250078558921814, "ttft_ms": 13.2552080322057, "decode_ms": 133.46712500788271, "receiver_prefill_ms": 248.80941689480096}, "total_wall_ms": 399.00816697627306, "ttft_ms": 13.2552080322057, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "24,521,113", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "1acde7a06be99d32a938c209eb287ec81faba3dedb13cc43c3a7e23397c698ee", "output_hash": "01fb8583197c19b3b9223fb6ed8963834d1675df14160ea6a5e8f145356e5145"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "37df54d3ceeb6c82736ab108", "attempt_id": "d8ffdf14fd66", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:57:00.840048Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 2267, "message_tokens": 2177, "message_utf8_bytes": 8286, "cache_tensor_bytes": null, "serialized_payload_bytes": 8286, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.033250078558921814, "ttft_ms": 13.327292050234973, "decode_ms": 13.290417031385005, "receiver_prefill_ms": 248.95179201848805}, "total_wall_ms": 279.10120808519423, "ttft_ms": 13.327292050234973, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "1acde7a06be99d32a938c209eb287ec81faba3dedb13cc43c3a7e23397c698ee", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "899e287efc383f86fbcfe57f", "attempt_id": "7bb8f7c959bf", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:57:01.133973Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 2268, "message_tokens": 2177, "message_utf8_bytes": 8286, "cache_tensor_bytes": null, "serialized_payload_bytes": 8286, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.033250078558921814, "ttft_ms": 13.391374959610403, "decode_ms": 13.366708997637033, "receiver_prefill_ms": 249.7671659803018}, "total_wall_ms": 279.9459161469713, "ttft_ms": 13.391374959610403, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "1acde7a06be99d32a938c209eb287ec81faba3dedb13cc43c3a7e23397c698ee", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "ecd91b69c2bd88b7726e02b0", "attempt_id": "b8053350a542", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:57:15.384200Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 1124, "message_tokens": 1036, "message_utf8_bytes": 2705, "cache_tensor_bytes": null, "serialized_payload_bytes": 2705, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13955.451374989934, "ttft_ms": 13.263124972581863, "decode_ms": 120.26724999304861, "receiver_prefill_ms": 121.62254203576595}, "total_wall_ms": 14212.495583924465, "ttft_ms": 13.263124972581863, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5,863,814", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "6d003a0ddf62889a583b0cd29bea24f90388af46dcf52134a5026edf2a229885", "output_hash": "f1722ba741a0ef5b3ac45cce5ffc57bea78f7eb0a07f17be098ba98a15ed3374"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "f09fe650ae7d1b4b3a9419e0", "attempt_id": "7b4733399f28", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:57:15.659134Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 1126, "message_tokens": 1036, "message_utf8_bytes": 2705, "cache_tensor_bytes": null, "serialized_payload_bytes": 2705, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13955.451374989934, "ttft_ms": 13.319082907401025, "decode_ms": 132.13191705290228, "receiver_prefill_ms": 121.26954202540219}, "total_wall_ms": 14223.995166947134, "ttft_ms": 13.319082907401025, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "NK-95Z-2165", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "6d003a0ddf62889a583b0cd29bea24f90388af46dcf52134a5026edf2a229885", "output_hash": "bd8f809b02304c6b9dc7c1e116ecccc2328c976a5c071e89a94d661b48e73aa7"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "977d3a382a65eb238eef4564", "attempt_id": "b5f460d0ac8f", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:57:15.942612Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 1133, "message_tokens": 1036, "message_utf8_bytes": 2705, "cache_tensor_bytes": null, "serialized_payload_bytes": 2705, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13955.451374989934, "ttft_ms": 13.252916047349572, "decode_ms": 79.88645799923688, "receiver_prefill_ms": 121.26983294729143}, "total_wall_ms": 14171.65841604583, "ttft_ms": 13.252916047349572, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "848090", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "6d003a0ddf62889a583b0cd29bea24f90388af46dcf52134a5026edf2a229885", "output_hash": "7db404962599d101fe0ffc53faae3729f0689e9529de8fe7e05cd115f1d2e1eb"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "271ee492e9a6dbdc2fd59024", "attempt_id": "eac56dc3a8f8", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:57:16.173945Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 1126, "message_tokens": 1036, "message_utf8_bytes": 2705, "cache_tensor_bytes": null, "serialized_payload_bytes": 2705, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13955.451374989934, "ttft_ms": 13.34929105360061, "decode_ms": 81.19833306409419, "receiver_prefill_ms": 125.21991599351168}, "total_wall_ms": 14177.089916891418, "ttft_ms": 13.34929105360061, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "848090", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "6d003a0ddf62889a583b0cd29bea24f90388af46dcf52134a5026edf2a229885", "output_hash": "7db404962599d101fe0ffc53faae3729f0689e9529de8fe7e05cd115f1d2e1eb"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "9bebd8a7a88866c46bbe6d91", "attempt_id": "fe2917154017", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:57:16.410199Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 1133, "message_tokens": 1036, "message_utf8_bytes": 2705, "cache_tensor_bytes": null, "serialized_payload_bytes": 2705, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13955.451374989934, "ttft_ms": 13.713291962631047, "decode_ms": 80.32979106064886, "receiver_prefill_ms": 137.8688330296427}, "total_wall_ms": 14189.201166969724, "ttft_ms": 13.713291962631047, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Gamtiszen-2", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "6d003a0ddf62889a583b0cd29bea24f90388af46dcf52134a5026edf2a229885", "output_hash": "32e3af428886434823d82102e1c482765eafdf6077fe8d3fba5752ec5e29826c"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "1e5bc9c9d203597b94287e8b", "attempt_id": "9fe70196bdd8", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:57:16.658361Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 1127, "message_tokens": 1036, "message_utf8_bytes": 2705, "cache_tensor_bytes": null, "serialized_payload_bytes": 2705, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13955.451374989934, "ttft_ms": 13.631792040541768, "decode_ms": 134.56395897082984, "receiver_prefill_ms": 142.6251670345664}, "total_wall_ms": 14248.060415964574, "ttft_ms": 13.631792040541768, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "15,647,775", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "6d003a0ddf62889a583b0cd29bea24f90388af46dcf52134a5026edf2a229885", "output_hash": "ad49154af6fd5cddbf2ef99540631d1caa98351d7ce0da191578b6e2459195ad"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "f9383d8e0e5e6e21f2a8dfd2", "attempt_id": "371c2ddef671", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:57:16.965264Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 1128, "message_tokens": 1036, "message_utf8_bytes": 2705, "cache_tensor_bytes": null, "serialized_payload_bytes": 2705, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13955.451374989934, "ttft_ms": 13.521249988116324, "decode_ms": 13.631540932692587, "receiver_prefill_ms": 136.61412498913705}, "total_wall_ms": 14121.080665965565, "ttft_ms": 13.521249988116324, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "6d003a0ddf62889a583b0cd29bea24f90388af46dcf52134a5026edf2a229885", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "771e31b5e82a4aea434d3ab4", "attempt_id": "4c055bb64589", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:57:17.145214Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 1129, "message_tokens": 1036, "message_utf8_bytes": 2705, "cache_tensor_bytes": null, "serialized_payload_bytes": 2705, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13955.451374989934, "ttft_ms": 13.746375101618469, "decode_ms": 13.576625031419098, "receiver_prefill_ms": 142.41345808841288}, "total_wall_ms": 14126.96979206521, "ttft_ms": 13.746375101618469, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "6d003a0ddf62889a583b0cd29bea24f90388af46dcf52134a5026edf2a229885", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "17664f12423af786416992ca", "attempt_id": "7c0f5ed11f32", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:57:31.318785Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 17}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 745, "message_tokens": 654, "message_utf8_bytes": 1802, "cache_tensor_bytes": null, "serialized_payload_bytes": 1802, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13987.811040948145, "ttft_ms": 13.392125023528934, "decode_ms": 13.21279106196016, "receiver_prefill_ms": 81.60441706422716}, "total_wall_ms": 14097.452832967974, "ttft_ms": 13.392125023528934, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "2ff665d4a3861c643fd3529595fadebd9e496d58804e3076811a27219a29fc3b", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "7026842716bce9a59bb7695f", "attempt_id": "7ca662b80189", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:57:31.448288Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 17}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 747, "message_tokens": 654, "message_utf8_bytes": 1802, "cache_tensor_bytes": null, "serialized_payload_bytes": 1802, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13987.811040948145, "ttft_ms": 13.118083006702363, "decode_ms": 13.009457965381444, "receiver_prefill_ms": 81.47679199464619}, "total_wall_ms": 14096.764915972017, "ttft_ms": 13.118083006702363, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "2ff665d4a3861c643fd3529595fadebd9e496d58804e3076811a27219a29fc3b", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c69496349a06f3910d1cc75b", "attempt_id": "4d4bfb4cf369", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:57:31.572362Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 17}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 754, "message_tokens": 654, "message_utf8_bytes": 1802, "cache_tensor_bytes": null, "serialized_payload_bytes": 1802, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13987.811040948145, "ttft_ms": 13.639333075843751, "decode_ms": 93.20912498515099, "receiver_prefill_ms": 81.42429206054658}, "total_wall_ms": 14177.486123982817, "ttft_ms": 13.639333075843751, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1065436", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "2ff665d4a3861c643fd3529595fadebd9e496d58804e3076811a27219a29fc3b", "output_hash": "09d8ffb65f02ffb967cf6e17cd51363071e72e6b8c774e5a5a670218f35a0c7a"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "94300bbc242498e6a0cbe093", "attempt_id": "81cbdf666f05", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:57:31.777063Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 17}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 747, "message_tokens": 654, "message_utf8_bytes": 1802, "cache_tensor_bytes": null, "serialized_payload_bytes": 1802, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13987.811040948145, "ttft_ms": 13.238750048913062, "decode_ms": 119.50620799325407, "receiver_prefill_ms": 80.69804206024855}, "total_wall_ms": 14202.605332946405, "ttft_ms": 13.238750048913062, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1,065,436", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "2ff665d4a3861c643fd3529595fadebd9e496d58804e3076811a27219a29fc3b", "output_hash": "58942f3111b469f049606f91be574a41a31c3fac8d5f41c17f49f21e7b507275"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "6b316b5508e31f83376f9d04", "attempt_id": "cd718046c3d6", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:57:32.006357Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 17}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 754, "message_tokens": 654, "message_utf8_bytes": 1802, "cache_tensor_bytes": null, "serialized_payload_bytes": 1802, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13987.811040948145, "ttft_ms": 13.610166031867266, "decode_ms": 78.86841706931591, "receiver_prefill_ms": 80.98991599399596}, "total_wall_ms": 14162.640373921022, "ttft_ms": 13.610166031867266, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Gamtiszen-2", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "2ff665d4a3861c643fd3529595fadebd9e496d58804e3076811a27219a29fc3b", "output_hash": "32e3af428886434823d82102e1c482765eafdf6077fe8d3fba5752ec5e29826c"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "ad16073fbc7ec56f7233cdc8", "attempt_id": "470cfd5eb376", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:57:32.195307Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 17}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 748, "message_tokens": 654, "message_utf8_bytes": 1802, "cache_tensor_bytes": null, "serialized_payload_bytes": 1802, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13987.811040948145, "ttft_ms": 13.2673749467358, "decode_ms": 92.79804094694555, "receiver_prefill_ms": 80.72608395013958}, "total_wall_ms": 14175.936207873747, "ttft_ms": 13.2673749467358, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1904227", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "2ff665d4a3861c643fd3529595fadebd9e496d58804e3076811a27219a29fc3b", "output_hash": "090088d1abac6f00e8a5180c91f2694f951035002ff5c23c5316b21dacb13129"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "1e614281b2b7dd383dd5f50c", "attempt_id": "e77811168e91", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:57:32.400303Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 17}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 749, "message_tokens": 654, "message_utf8_bytes": 1802, "cache_tensor_bytes": null, "serialized_payload_bytes": 1802, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13987.811040948145, "ttft_ms": 13.32866691518575, "decode_ms": 13.212374993599951, "receiver_prefill_ms": 80.68941591773182}, "total_wall_ms": 14096.456583007239, "ttft_ms": 13.32866691518575, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "2ff665d4a3861c643fd3529595fadebd9e496d58804e3076811a27219a29fc3b", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "ea3e23e8ef7d1dea65bcb26d", "attempt_id": "4c305048ff04", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:57:32.523495Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 17}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 750, "message_tokens": 654, "message_utf8_bytes": 1802, "cache_tensor_bytes": null, "serialized_payload_bytes": 1802, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13987.811040948145, "ttft_ms": 13.274834025651217, "decode_ms": 13.22729210369289, "receiver_prefill_ms": 81.71662501990795}, "total_wall_ms": 14097.426873981021, "ttft_ms": 13.274834025651217, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "2ff665d4a3861c643fd3529595fadebd9e496d58804e3076811a27219a29fc3b", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "7d7c03b5ebe2458884f65e3b", "attempt_id": "ac96dfeb158c", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:57:46.598769Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 1350, "message_tokens": 1260, "message_utf8_bytes": 3302, "cache_tensor_bytes": null, "serialized_payload_bytes": 3302, "side_information_bytes": 0, "fetched_bytes": 1482, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13948.762041982263, "ttft_ms": 13.435167027637362, "decode_ms": 13.125792029313743, "receiver_prefill_ms": 144.739874987863}, "total_wall_ms": 14122.174000949599, "ttft_ms": 13.435167027637362, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "89f5673f685ae9f2e43d18218ee15719605160a10e3bc586bf859e728f223871", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "7db6e8e39275bb9a718b10e6", "attempt_id": "e6c01cd60ebd", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:57:46.787043Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 1326, "message_tokens": 1234, "message_utf8_bytes": 3240, "cache_tensor_bytes": null, "serialized_payload_bytes": 3240, "side_information_bytes": 0, "fetched_bytes": 1442, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13948.762041982263, "ttft_ms": 13.69995903223753, "decode_ms": 13.630624976940453, "receiver_prefill_ms": 143.62862496636808}, "total_wall_ms": 14121.765249874443, "ttft_ms": 13.69995903223753, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "f11a1a265309ca9bfe4e945b7f164dfd6e8dfa355cb261df47ec7b2d54689109", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "e67251e6aa7b4a9e2cc5cffb", "attempt_id": "1dc95ddbf54e", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:57:46.974432Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 1359, "message_tokens": 1260, "message_utf8_bytes": 3302, "cache_tensor_bytes": null, "serialized_payload_bytes": 3302, "side_information_bytes": 0, "fetched_bytes": 1482, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13948.762041982263, "ttft_ms": 13.38816701900214, "decode_ms": 80.63104201573879, "receiver_prefill_ms": 152.46666700113565}, "total_wall_ms": 14197.35083298292, "ttft_ms": 13.38816701900214, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "911925", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "89f5673f685ae9f2e43d18218ee15719605160a10e3bc586bf859e728f223871", "output_hash": "679a540dc4739f0ace623448b8fd925fb234f554838d7e87da223228ee81b112"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "291076ef92a0d5f8bfcb191f", "attempt_id": "7e749e0e30b4", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:57:47.237458Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 1352, "message_tokens": 1260, "message_utf8_bytes": 3302, "cache_tensor_bytes": null, "serialized_payload_bytes": 3302, "side_information_bytes": 0, "fetched_bytes": 1482, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13948.762041982263, "ttft_ms": 13.962415978312492, "decode_ms": 81.0849160188809, "receiver_prefill_ms": 173.8602500408888}, "total_wall_ms": 14219.718333915807, "ttft_ms": 13.962415978312492, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "911925", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "89f5673f685ae9f2e43d18218ee15719605160a10e3bc586bf859e728f223871", "output_hash": "679a540dc4739f0ace623448b8fd925fb234f554838d7e87da223228ee81b112"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b1209026783c4c262ceebba8", "attempt_id": "1a6b84f3163a", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:57:47.523116Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 1344, "message_tokens": 1245, "message_utf8_bytes": 3262, "cache_tensor_bytes": null, "serialized_payload_bytes": 3262, "side_information_bytes": 0, "fetched_bytes": 1471, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13948.762041982263, "ttft_ms": 13.729959027841687, "decode_ms": 80.80070791766047, "receiver_prefill_ms": 170.40199995972216}, "total_wall_ms": 14215.742166968994, "ttft_ms": 13.729959027841687, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Gamtiszen-2", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "7b1cc3815ad150bf219a0c9ab0d7542604908ff524d178575b2aba171f39def6", "output_hash": "32e3af428886434823d82102e1c482765eafdf6077fe8d3fba5752ec5e29826c"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "30e6348f995d42c1a8405a13", "attempt_id": "aa9479188a5f", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:57:47.804341Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 1353, "message_tokens": 1260, "message_utf8_bytes": 3302, "cache_tensor_bytes": null, "serialized_payload_bytes": 3302, "side_information_bytes": 0, "fetched_bytes": 1482, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13948.762041982263, "ttft_ms": 13.576042023487389, "decode_ms": 134.45083401165903, "receiver_prefill_ms": 170.13574996963143}, "total_wall_ms": 14269.00970900897, "ttft_ms": 13.576042023487389, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "19,220,405", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "89f5673f685ae9f2e43d18218ee15719605160a10e3bc586bf859e728f223871", "output_hash": "0f3eec909a84d46db9c779728a9b606b28bb1d7260cddda4d39193b7f7f596f3"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "61eaf6b6e0bc7220e4171fb4", "attempt_id": "ac81b574c27c", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:57:48.138889Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 1334, "message_tokens": 1240, "message_utf8_bytes": 3309, "cache_tensor_bytes": null, "serialized_payload_bytes": 3309, "side_information_bytes": 0, "fetched_bytes": 1491, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13948.762041982263, "ttft_ms": 13.542542001232505, "decode_ms": 13.437791960313916, "receiver_prefill_ms": 160.214125062339}, "total_wall_ms": 14138.049417058937, "ttft_ms": 13.542542001232505, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "3708cb808d1b0d12aeb4d624f32bc8b82b405ccf4d18e077d6b279003ccbcad2", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "9bd0c14fea1fc1fef0f07d0a", "attempt_id": "09d8686676f4", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:57:48.342237Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 1361, "message_tokens": 1266, "message_utf8_bytes": 3339, "cache_tensor_bytes": null, "serialized_payload_bytes": 3339, "side_information_bytes": 0, "fetched_bytes": 1531, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13948.762041982263, "ttft_ms": 13.658124953508377, "decode_ms": 13.557583908550441, "receiver_prefill_ms": 170.11095897760242}, "total_wall_ms": 14148.130999994464, "ttft_ms": 13.658124953508377, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a5fed066a9d6c05f88c63db51b898727366e25abdf699f987a57c0ccb09c4087", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "7a54a0b8b4b864e2d2a29797", "attempt_id": "3231e0972887", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:57:48.863644Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 83128320}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 44, "message_tokens": null, "message_utf8_bytes": 8260, "cache_tensor_bytes": 83128320, "serialized_payload_bytes": 83128320, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 304.95154100935906, "ttft_ms": 33.541333046741784, "decode_ms": 125.77758403494954, "receiver_prefill_ms": 0.0}, "total_wall_ms": 468.5222489060834, "ttft_ms": 33.541333046741784, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5,863,814", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "f1722ba741a0ef5b3ac45cce5ffc57bea78f7eb0a07f17be098ba98a15ed3374"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "22f6a739c2617f9efb5d48b1", "attempt_id": "b63c25421730", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:57:49.049430Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 83128320}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 46, "message_tokens": null, "message_utf8_bytes": 8260, "cache_tensor_bytes": 83128320, "serialized_payload_bytes": 83128320, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 304.95154100935906, "ttft_ms": 29.150625108741224, "decode_ms": 138.3447089465335, "receiver_prefill_ms": 0.0}, "total_wall_ms": 475.32095795031637, "ttft_ms": 29.150625108741224, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "NK-95Z-2165", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "bd8f809b02304c6b9dc7c1e116ecccc2328c976a5c071e89a94d661b48e73aa7"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "0a6c3a8a92da06720f5b567b", "attempt_id": "aed04af7a3ca", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:57:49.236741Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 83128320}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 53, "message_tokens": null, "message_utf8_bytes": 8260, "cache_tensor_bytes": 83128320, "serialized_payload_bytes": 83128320, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 304.95154100935906, "ttft_ms": 26.94612496998161, "decode_ms": 123.3197080437094, "receiver_prefill_ms": 0.0}, "total_wall_ms": 458.0844580195844, "ttft_ms": 26.94612496998161, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1,065,436", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "58942f3111b469f049606f91be574a41a31c3fac8d5f41c17f49f21e7b507275"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "717282c5650120b4e1292b8d", "attempt_id": "681522e2fe6c", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:57:49.406625Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 83128320}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 46, "message_tokens": null, "message_utf8_bytes": 8260, "cache_tensor_bytes": 83128320, "serialized_payload_bytes": 83128320, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 304.95154100935906, "ttft_ms": 27.248375001363456, "decode_ms": 13.712167041376233, "receiver_prefill_ms": 0.0}, "total_wall_ms": 349.8201660113409, "ttft_ms": 27.248375001363456, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "9b74f6e3c9c3318eaa83dc17", "attempt_id": "80532e1b80ec", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:57:49.465531Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 83128320}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 53, "message_tokens": null, "message_utf8_bytes": 8260, "cache_tensor_bytes": 83128320, "serialized_payload_bytes": 83128320, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 304.95154100935906, "ttft_ms": 26.247708941809833, "decode_ms": 83.68799998424947, "receiver_prefill_ms": 0.0}, "total_wall_ms": 418.67612407077104, "ttft_ms": 26.247708941809833, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Gamtiszen-2", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "32e3af428886434823d82102e1c482765eafdf6077fe8d3fba5752ec5e29826c"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "8b08b3d1e6169377b9b66a7e", "attempt_id": "bd37065a2a43", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:57:49.593289Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 83128320}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 8260, "cache_tensor_bytes": 83128320, "serialized_payload_bytes": 83128320, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 304.95154100935906, "ttft_ms": 27.542124968022108, "decode_ms": 140.61591611243784, "receiver_prefill_ms": 0.0}, "total_wall_ms": 476.5418750466779, "ttft_ms": 27.542124968022108, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "17,123,004", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "771a05561784dc1565b6394f21cf903e5a7f529d18d44ac10055aa9a9d1efea4"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "532c3aa8f7adb4555ab41814", "attempt_id": "26d093cf6717", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:57:49.780064Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 83128320}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 48, "message_tokens": null, "message_utf8_bytes": 8260, "cache_tensor_bytes": 83128320, "serialized_payload_bytes": 83128320, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 304.95154100935906, "ttft_ms": 27.469833032228053, "decode_ms": 14.043167000636458, "receiver_prefill_ms": 0.0}, "total_wall_ms": 350.36391601897776, "ttft_ms": 27.469833032228053, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "ce40f0785d1431edc92c4bea", "attempt_id": "503643088c8f", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:57:49.839856Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 83128320}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 49, "message_tokens": null, "message_utf8_bytes": 8260, "cache_tensor_bytes": 83128320, "serialized_payload_bytes": 83128320, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 304.95154100935906, "ttft_ms": 25.339209008961916, "decode_ms": 13.812417048029602, "receiver_prefill_ms": 0.0}, "total_wall_ms": 347.9434580076486, "ttft_ms": 25.339209008961916, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "774add3f0f0ddd5bac22ed59", "attempt_id": "c9a48af8572a", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:57:49.898285Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 2377, "message_tokens": 2291, "message_utf8_bytes": 8790, "cache_tensor_bytes": null, "serialized_payload_bytes": 8790, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.02487492747604847, "ttft_ms": 13.755874941125512, "decode_ms": 121.56745803076774, "receiver_prefill_ms": 268.63691699691117}, "total_wall_ms": 407.60879102163017, "ttft_ms": 13.755874941125512, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5,863,814", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "7e449bb35238caf0ecc0f02a71cf642a25624204d8deacc60c8133a971bdf010", "output_hash": "f1722ba741a0ef5b3ac45cce5ffc57bea78f7eb0a07f17be098ba98a15ed3374"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "20dc406b09d870a8eb54b3b2", "attempt_id": "f8a9bf1228dc", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:57:50.319987Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 2379, "message_tokens": 2291, "message_utf8_bytes": 8790, "cache_tensor_bytes": null, "serialized_payload_bytes": 8790, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.02487492747604847, "ttft_ms": 13.512542005628347, "decode_ms": 134.69804101623595, "receiver_prefill_ms": 272.9770829901099}, "total_wall_ms": 424.7602499090135, "ttft_ms": 13.512542005628347, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "NK-95Z-2165", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "7e449bb35238caf0ecc0f02a71cf642a25624204d8deacc60c8133a971bdf010", "output_hash": "bd8f809b02304c6b9dc7c1e116ecccc2328c976a5c071e89a94d661b48e73aa7"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "fe79f5869739cfd2c6f34cbf", "attempt_id": "d03a360fd0fb", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:57:50.763860Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 2386, "message_tokens": 2291, "message_utf8_bytes": 8790, "cache_tensor_bytes": null, "serialized_payload_bytes": 8790, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.02487492747604847, "ttft_ms": 13.600458041764796, "decode_ms": 80.85825003217906, "receiver_prefill_ms": 277.75270806159824}, "total_wall_ms": 375.823708018288, "ttft_ms": 13.600458041764796, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "848090", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "7e449bb35238caf0ecc0f02a71cf642a25624204d8deacc60c8133a971bdf010", "output_hash": "7db404962599d101fe0ffc53faae3729f0689e9529de8fe7e05cd115f1d2e1eb"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "1c6401c767b149712af3dda5", "attempt_id": "67b078e3b322", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:57:51.156740Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 2379, "message_tokens": 2291, "message_utf8_bytes": 8790, "cache_tensor_bytes": null, "serialized_payload_bytes": 8790, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.02487492747604847, "ttft_ms": 14.020542032085359, "decode_ms": 94.6674169972539, "receiver_prefill_ms": 290.75041704345495}, "total_wall_ms": 403.1262499047443, "ttft_ms": 14.020542032085359, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "848,090", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "7e449bb35238caf0ecc0f02a71cf642a25624204d8deacc60c8133a971bdf010", "output_hash": "b5105383225daa63c1554bd08d37339a9b2d2a1066058fc15ab16cf75f05d898"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c8ddc3b00be8ab6065e16cea", "attempt_id": "7271ced4de78", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:57:51.574628Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 2386, "message_tokens": 2291, "message_utf8_bytes": 8790, "cache_tensor_bytes": null, "serialized_payload_bytes": 8790, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.02487492747604847, "ttft_ms": 13.662416953593493, "decode_ms": 81.51899999938905, "receiver_prefill_ms": 293.755958089605}, "total_wall_ms": 392.51637493725866, "ttft_ms": 13.662416953593493, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Gamtiszen-2", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "7e449bb35238caf0ecc0f02a71cf642a25624204d8deacc60c8133a971bdf010", "output_hash": "32e3af428886434823d82102e1c482765eafdf6077fe8d3fba5752ec5e29826c"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "43194f16b5906abfe24fa301", "attempt_id": "5c16b3e8d2b1", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:57:51.982427Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 2380, "message_tokens": 2291, "message_utf8_bytes": 8790, "cache_tensor_bytes": null, "serialized_payload_bytes": 8790, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.02487492747604847, "ttft_ms": 13.731041923165321, "decode_ms": 135.0337500916794, "receiver_prefill_ms": 295.1212910702452}, "total_wall_ms": 447.6203339872882, "ttft_ms": 13.731041923165321, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "35,528,789", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "7e449bb35238caf0ecc0f02a71cf642a25624204d8deacc60c8133a971bdf010", "output_hash": "09c62e54f01a16ad4592acbc72b207844cda0e543839bafa2975a3c86bca59be"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "4feef0f6ef5ecefa7cf08970", "attempt_id": "b7d07d4c686a", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:57:52.444425Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 2381, "message_tokens": 2291, "message_utf8_bytes": 8790, "cache_tensor_bytes": null, "serialized_payload_bytes": 8790, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.02487492747604847, "ttft_ms": 13.661624980159104, "decode_ms": 13.621458085253835, "receiver_prefill_ms": 299.1212500492111}, "total_wall_ms": 330.0826669437811, "ttft_ms": 13.661624980159104, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "7e449bb35238caf0ecc0f02a71cf642a25624204d8deacc60c8133a971bdf010", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a95f6f626baeb43aff8a36e0", "attempt_id": "452064caca83", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:57:52.788979Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 2382, "message_tokens": 2291, "message_utf8_bytes": 8790, "cache_tensor_bytes": null, "serialized_payload_bytes": 8790, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.02487492747604847, "ttft_ms": 13.861708925105631, "decode_ms": 13.602916966192424, "receiver_prefill_ms": 305.5283339926973}, "total_wall_ms": 336.64337487425655, "ttft_ms": 13.861708925105631, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "7e449bb35238caf0ecc0f02a71cf642a25624204d8deacc60c8133a971bdf010", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "1b13a80b0c6531d61b7e5426", "attempt_id": "1d64d9a2fd77", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:58:15.012758Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1611}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 1722, "message_tokens": 1635, "message_utf8_bytes": 5599, "cache_tensor_bytes": null, "serialized_payload_bytes": 5599, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 21872.222124948166, "ttft_ms": 13.63620802294463, "decode_ms": 93.22920790873468, "receiver_prefill_ms": 180.9904589317739}, "total_wall_ms": 22163.540291949175, "ttft_ms": 13.63620802294463, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5863814", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "f952ad8c7f1fbf5e8b60ca5f9ac097f3d873b364cd806a8c2e1d553857f14dee", "output_hash": "8a4efc21da7fd0aafdbb1f8f5fe9231f746cd4dd77afd1da76a2f1ed0b271b75"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "e450991714f13a197bf787a3", "attempt_id": "f45dc246d080", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:58:15.319296Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1611}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 1724, "message_tokens": 1635, "message_utf8_bytes": 5599, "cache_tensor_bytes": null, "serialized_payload_bytes": 5599, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 21872.222124948166, "ttft_ms": 13.52849998511374, "decode_ms": 134.31916700210422, "receiver_prefill_ms": 180.15141692012548}, "total_wall_ms": 22202.867790940218, "ttft_ms": 13.52849998511374, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "NK-95Z-2165", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "f952ad8c7f1fbf5e8b60ca5f9ac097f3d873b364cd806a8c2e1d553857f14dee", "output_hash": "bd8f809b02304c6b9dc7c1e116ecccc2328c976a5c071e89a94d661b48e73aa7"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "5170efdcfbcad351fa39ff84", "attempt_id": "8de5889c0ac8", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:58:15.664708Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1611}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 1731, "message_tokens": 1635, "message_utf8_bytes": 5599, "cache_tensor_bytes": null, "serialized_payload_bytes": 5599, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 21872.222124948166, "ttft_ms": 13.651250046677887, "decode_ms": 80.7345419889316, "receiver_prefill_ms": 193.0127500090748}, "total_wall_ms": 22162.29274997022, "ttft_ms": 13.651250046677887, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "848090", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "f952ad8c7f1fbf5e8b60ca5f9ac097f3d873b364cd806a8c2e1d553857f14dee", "output_hash": "7db404962599d101fe0ffc53faae3729f0689e9529de8fe7e05cd115f1d2e1eb"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "2456fd4c166de9c331d70c44", "attempt_id": "bbb3c987fd53", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:58:15.969734Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1611}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 1724, "message_tokens": 1635, "message_utf8_bytes": 5599, "cache_tensor_bytes": null, "serialized_payload_bytes": 5599, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 21872.222124948166, "ttft_ms": 13.979375013150275, "decode_ms": 95.18308297265321, "receiver_prefill_ms": 210.16925002913922}, "total_wall_ms": 22194.244416896254, "ttft_ms": 13.979375013150275, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "848,090", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "f952ad8c7f1fbf5e8b60ca5f9ac097f3d873b364cd806a8c2e1d553857f14dee", "output_hash": "b5105383225daa63c1554bd08d37339a9b2d2a1066058fc15ab16cf75f05d898"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "4c03778fd1d5abdc8f5049b2", "attempt_id": "66d7ce9c50d2", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:58:16.306720Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1611}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 1731, "message_tokens": 1635, "message_utf8_bytes": 5599, "cache_tensor_bytes": null, "serialized_payload_bytes": 5599, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 21872.222124948166, "ttft_ms": 13.910292065702379, "decode_ms": 81.54895901679993, "receiver_prefill_ms": 218.60941697377712}, "total_wall_ms": 22188.943249871954, "ttft_ms": 13.910292065702379, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Gamtiszen-2", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "f952ad8c7f1fbf5e8b60ca5f9ac097f3d873b364cd806a8c2e1d553857f14dee", "output_hash": "32e3af428886434823d82102e1c482765eafdf6077fe8d3fba5752ec5e29826c"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a1f72ea1611e870a922cea2f", "attempt_id": "028401c26f24", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:58:16.638867Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1611}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 1725, "message_tokens": 1635, "message_utf8_bytes": 5599, "cache_tensor_bytes": null, "serialized_payload_bytes": 5599, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 21872.222124948166, "ttft_ms": 13.883208972401917, "decode_ms": 134.64841700624675, "receiver_prefill_ms": 215.06470895837992}, "total_wall_ms": 22238.39987500105, "ttft_ms": 13.883208972401917, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "15,688,579", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "f952ad8c7f1fbf5e8b60ca5f9ac097f3d873b364cd806a8c2e1d553857f14dee", "output_hash": "15f7ad292e4f3975f5923f223cca8ff35838e353dcda8670c2a8f427d80bd67a"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "790bba0766448d1c761042f3", "attempt_id": "0f27cf7dc42c", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:58:17.019011Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1611}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 1726, "message_tokens": 1635, "message_utf8_bytes": 5599, "cache_tensor_bytes": null, "serialized_payload_bytes": 5599, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 21872.222124948166, "ttft_ms": 13.567666988819838, "decode_ms": 13.651833985932171, "receiver_prefill_ms": 205.42170794215053}, "total_wall_ms": 22107.47183393687, "ttft_ms": 13.567666988819838, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "f952ad8c7f1fbf5e8b60ca5f9ac097f3d873b364cd806a8c2e1d553857f14dee", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b22388597bbdad2c8931698c", "attempt_id": "7c18b138bc82", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:58:17.269071Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1611}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 1727, "message_tokens": 1635, "message_utf8_bytes": 5599, "cache_tensor_bytes": null, "serialized_payload_bytes": 5599, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 21872.222124948166, "ttft_ms": 13.81020899862051, "decode_ms": 13.558708014898002, "receiver_prefill_ms": 214.82566697522998}, "total_wall_ms": 22117.025958024897, "ttft_ms": 13.81020899862051, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "f952ad8c7f1fbf5e8b60ca5f9ac097f3d873b364cd806a8c2e1d553857f14dee", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "db99ec3f77964f019b00a42f", "attempt_id": "9f79c116dea8", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:58:41.693704Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 27}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 1103, "message_tokens": 1012, "message_utf8_bytes": 2669, "cache_tensor_bytes": null, "serialized_payload_bytes": 2669, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24165.432000067085, "ttft_ms": 13.455791980959475, "decode_ms": 13.061792007647455, "receiver_prefill_ms": 117.48612497467548}, "total_wall_ms": 24311.23470806051, "ttft_ms": 13.455791980959475, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "76f460e04cbd109e753bc16ba653bebc3fe10ef745a1d209025e4c9409989358", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "50805efa318964dd74f9f5dc", "attempt_id": "36c7748ffe26", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:58:41.854416Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 27}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 1105, "message_tokens": 1012, "message_utf8_bytes": 2669, "cache_tensor_bytes": null, "serialized_payload_bytes": 2669, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24165.432000067085, "ttft_ms": 13.43216688837856, "decode_ms": 132.32637499459088, "receiver_prefill_ms": 117.71129199769348}, "total_wall_ms": 24430.6267920183, "ttft_ms": 13.43216688837856, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "NK-95Z-2165", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "76f460e04cbd109e753bc16ba653bebc3fe10ef745a1d209025e4c9409989358", "output_hash": "bd8f809b02304c6b9dc7c1e116ecccc2328c976a5c071e89a94d661b48e73aa7"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "d0e3eda5726eec0c0424ad42", "attempt_id": "697849bdb02a", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:58:42.135136Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 27}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 1112, "message_tokens": 1012, "message_utf8_bytes": 2669, "cache_tensor_bytes": null, "serialized_payload_bytes": 2669, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24165.432000067085, "ttft_ms": 13.623374979943037, "decode_ms": 93.30649999901652, "receiver_prefill_ms": 118.40633291285485}, "total_wall_ms": 24392.578250030056, "ttft_ms": 13.623374979943037, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "911,925", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "76f460e04cbd109e753bc16ba653bebc3fe10ef745a1d209025e4c9409989358", "output_hash": "86be78720fc4b79f93e603e7d9e4861fcd25102612f73209fffc17d787056007"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "7761fe3b2a060057dca8330a", "attempt_id": "660a058361cf", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:58:42.377064Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 27}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 1105, "message_tokens": 1012, "message_utf8_bytes": 2669, "cache_tensor_bytes": null, "serialized_payload_bytes": 2669, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24165.432000067085, "ttft_ms": 13.286959030665457, "decode_ms": 92.52045804169029, "receiver_prefill_ms": 117.58962506428361}, "total_wall_ms": 24390.573333017528, "ttft_ms": 13.286959030665457, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "911,925", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "76f460e04cbd109e753bc16ba653bebc3fe10ef745a1d209025e4c9409989358", "output_hash": "86be78720fc4b79f93e603e7d9e4861fcd25102612f73209fffc17d787056007"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "6eeae41729d720d0240dd1e4", "attempt_id": "c9489cc4166c", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:58:42.618685Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 27}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 1112, "message_tokens": 1012, "message_utf8_bytes": 2669, "cache_tensor_bytes": null, "serialized_payload_bytes": 2669, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24165.432000067085, "ttft_ms": 13.72904097661376, "decode_ms": 81.2510410323739, "receiver_prefill_ms": 120.95958297140896}, "total_wall_ms": 24383.126000058837, "ttft_ms": 13.72904097661376, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Gamtiszen-2", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "76f460e04cbd109e753bc16ba653bebc3fe10ef745a1d209025e4c9409989358", "output_hash": "32e3af428886434823d82102e1c482765eafdf6077fe8d3fba5752ec5e29826c"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "80f1ac54cb26b479b2bd84cc", "attempt_id": "7f9552102766", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:58:42.850725Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 27}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 1106, "message_tokens": 1012, "message_utf8_bytes": 2669, "cache_tensor_bytes": null, "serialized_payload_bytes": 2669, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24165.432000067085, "ttft_ms": 14.646749943494797, "decode_ms": 95.44420905876905, "receiver_prefill_ms": 903.3094589831308}, "total_wall_ms": 25180.61516713351, "ttft_ms": 14.646749943494797, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "3419675", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "76f460e04cbd109e753bc16ba653bebc3fe10ef745a1d209025e4c9409989358", "output_hash": "6ff4683836956b23bf9fa006232365c7ad2e84db1352894b1c598c45f83e0196"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "2e974f258d8a3a6aa7611d67", "attempt_id": "9bccb1c74b34", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:58:43.880817Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 27}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 1107, "message_tokens": 1012, "message_utf8_bytes": 2669, "cache_tensor_bytes": null, "serialized_payload_bytes": 2669, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24165.432000067085, "ttft_ms": 13.573208008892834, "decode_ms": 14.137333957478404, "receiver_prefill_ms": 535.0551250157878}, "total_wall_ms": 24730.33741710242, "ttft_ms": 13.573208008892834, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "76f460e04cbd109e753bc16ba653bebc3fe10ef745a1d209025e4c9409989358", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "d6429b45c90602d9a5c83f9e", "attempt_id": "9411789d63de", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:58:44.460815Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 27}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 1108, "message_tokens": 1012, "message_utf8_bytes": 2669, "cache_tensor_bytes": null, "serialized_payload_bytes": 2669, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24165.432000067085, "ttft_ms": 13.929166016168892, "decode_ms": 13.873041956685483, "receiver_prefill_ms": 230.29745800886303}, "total_wall_ms": 24427.633417071775, "ttft_ms": 13.929166016168892, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "76f460e04cbd109e753bc16ba653bebc3fe10ef745a1d209025e4c9409989358", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "1694b7ba9cc366dd0334109f", "attempt_id": "31b6152b449c", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:59:09.317434Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 2533, "message_tokens": 2443, "message_utf8_bytes": 6389, "cache_tensor_bytes": null, "serialized_payload_bytes": 6389, "side_information_bytes": 0, "fetched_bytes": 3454, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24308.74562507961, "ttft_ms": 13.610958936624229, "decode_ms": 13.280584011226892, "receiver_prefill_ms": 331.5474579576403}, "total_wall_ms": 24671.791708096862, "ttft_ms": 13.610958936624229, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "7c79b5408c53acd013540f7ac081a51a2c6cc8852897ca40b9d7231c67709e12", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "7d675e3d6c3ec4da98b2cc58", "attempt_id": "bfe4e3ac4420", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:59:09.695155Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 2469, "message_tokens": 2377, "message_utf8_bytes": 6171, "cache_tensor_bytes": null, "serialized_payload_bytes": 6171, "side_information_bytes": 0, "fetched_bytes": 3232, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24308.74562507961, "ttft_ms": 14.623625087551773, "decode_ms": 138.8584580272436, "receiver_prefill_ms": 290.99837492685765}, "total_wall_ms": 24757.072542095557, "ttft_ms": 14.623625087551773, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "NK-95Z-2165", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a897321f1e88079c6dddfc35c11c7bfe8e9279e99b471307b14770d2b4468e7b", "output_hash": "bd8f809b02304c6b9dc7c1e116ecccc2328c976a5c071e89a94d661b48e73aa7"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "f0c73f39040b55a41d66a0de", "attempt_id": "8f0e7e14380d", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:59:10.158379Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 2542, "message_tokens": 2443, "message_utf8_bytes": 6389, "cache_tensor_bytes": null, "serialized_payload_bytes": 6389, "side_information_bytes": 0, "fetched_bytes": 3454, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24308.74562507961, "ttft_ms": 13.88329197652638, "decode_ms": 82.19116693362594, "receiver_prefill_ms": 436.16191600449383}, "total_wall_ms": 24844.361875089817, "ttft_ms": 13.88329197652638, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "911925", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "15405b4be3fa9bef155dd9e33e6222e58ee97e6f0acc44d0f354c59f928f1d8f", "output_hash": "679a540dc4739f0ace623448b8fd925fb234f554838d7e87da223228ee81b112"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "137a8950093241a9eec4ab6e", "attempt_id": "5ecf791979ca", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:59:10.708992Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 2535, "message_tokens": 2443, "message_utf8_bytes": 6389, "cache_tensor_bytes": null, "serialized_payload_bytes": 6389, "side_information_bytes": 0, "fetched_bytes": 3454, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24308.74562507961, "ttft_ms": 13.553541037254035, "decode_ms": 80.60237509198487, "receiver_prefill_ms": 317.18637503217906}, "total_wall_ms": 24723.52300013881, "ttft_ms": 13.553541037254035, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "911925", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "cbf20bf915fa4c467baefc2500fe9fe2f14a067175e76ba6f4d7dc377539da04", "output_hash": "679a540dc4739f0ace623448b8fd925fb234f554838d7e87da223228ee81b112"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "2d4c4c35d498683f4480934e", "attempt_id": "2dc2bd0f7378", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:59:11.139838Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 2522, "message_tokens": 2423, "message_utf8_bytes": 6320, "cache_tensor_bytes": null, "serialized_payload_bytes": 6320, "side_information_bytes": 0, "fetched_bytes": 3394, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24308.74562507961, "ttft_ms": 14.147624955512583, "decode_ms": 68.85270797647536, "receiver_prefill_ms": 340.82304197363555}, "total_wall_ms": 24735.801084083505, "ttft_ms": 14.147624955512583, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Brigam-1", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "9d48d7a359d6481166d1492a71032b694aee30294f70f66f3ee5ff1925cc2b3f", "output_hash": "cbeedb5efb33fd78c3c90ee9a3fd8829f55bdc158719dbd6bd04a8179d5b3aef"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b6c4889e73ed4451e587b768", "attempt_id": "d6178b03b485", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:59:11.581216Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 2536, "message_tokens": 2443, "message_utf8_bytes": 6389, "cache_tensor_bytes": null, "serialized_payload_bytes": 6389, "side_information_bytes": 0, "fetched_bytes": 3454, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24308.74562507961, "ttft_ms": 13.89391697011888, "decode_ms": 134.70508391037583, "receiver_prefill_ms": 306.26204202417284}, "total_wall_ms": 24766.898792120628, "ttft_ms": 13.89391697011888, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "30,219,805", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "963acab62e68ac4ef552f9c0f90219787a0be6145465d6d901032fcd2c76d1e2", "output_hash": "9c2c1ecce3e8904b1c70c8aea4551e92a7ab6333f6f7259eb3410623989488b5"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "ff1f0186b95bb24e09697992", "attempt_id": "2275c028ff7b", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:59:12.054409Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 2514, "message_tokens": 2420, "message_utf8_bytes": 6370, "cache_tensor_bytes": null, "serialized_payload_bytes": 6370, "side_information_bytes": 0, "fetched_bytes": 3409, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24308.74562507961, "ttft_ms": 13.835957972332835, "decode_ms": 13.525542104616761, "receiver_prefill_ms": 293.3732090750709}, "total_wall_ms": 24632.75175006129, "ttft_ms": 13.835957972332835, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "027c585fbbfb35b10989f0c3ca2d71331660cc9bcdc91b7649c18f180ea46063", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "7479eb701c8eb8818ca12888", "attempt_id": "d3d51f533ed0", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:59:12.392742Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 2538, "message_tokens": 2443, "message_utf8_bytes": 6389, "cache_tensor_bytes": null, "serialized_payload_bytes": 6389, "side_information_bytes": 0, "fetched_bytes": 3454, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24308.74562507961, "ttft_ms": 13.916833093389869, "decode_ms": 14.073624974116683, "receiver_prefill_ms": 344.7994999587536}, "total_wall_ms": 24684.875125065446, "ttft_ms": 13.916833093389869, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "56ca633f374fbecf06d6774628d59b60a2b70240b3d18e1faad8d6a9b9e9de30", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "573080cfe76011a12577dd5d", "attempt_id": "844f1caf8412", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:59:13.302587Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 90206208}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 44, "message_tokens": null, "message_utf8_bytes": 8736, "cache_tensor_bytes": 90206208, "serialized_payload_bytes": 90206208, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 519.2071249475703, "ttft_ms": 41.60741704981774, "decode_ms": 131.803167052567, "receiver_prefill_ms": 0.0}, "total_wall_ms": 696.4418329298496, "ttft_ms": 41.60741704981774, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5,863,814", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "f1722ba741a0ef5b3ac45cce5ffc57bea78f7eb0a07f17be098ba98a15ed3374"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "8cb2c093aeed215598f6c527", "attempt_id": "663770a77bc3", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:59:13.496962Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 90206208}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 46, "message_tokens": null, "message_utf8_bytes": 8736, "cache_tensor_bytes": 90206208, "serialized_payload_bytes": 90206208, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 519.2071249475703, "ttft_ms": 29.238375020213425, "decode_ms": 138.2906250655651, "receiver_prefill_ms": 0.0}, "total_wall_ms": 690.0612079771236, "ttft_ms": 29.238375020213425, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "NK-95Z-2165", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "bd8f809b02304c6b9dc7c1e116ecccc2328c976a5c071e89a94d661b48e73aa7"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "d0f2249c0e7d6788f3518b31", "attempt_id": "1cf5bc9f948d", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:59:13.682867Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 90206208}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 53, "message_tokens": null, "message_utf8_bytes": 8736, "cache_tensor_bytes": 90206208, "serialized_payload_bytes": 90206208, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 519.2071249475703, "ttft_ms": 28.862208011560142, "decode_ms": 97.03941701445729, "receiver_prefill_ms": 0.0}, "total_wall_ms": 647.306416882202, "ttft_ms": 28.862208011560142, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "848,090", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b5105383225daa63c1554bd08d37339a9b2d2a1066058fc15ab16cf75f05d898"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "bab87b50debd1f1bd38aae26", "attempt_id": "29aa2894c0dd", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:59:13.825963Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 90206208}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 46, "message_tokens": null, "message_utf8_bytes": 8736, "cache_tensor_bytes": 90206208, "serialized_payload_bytes": 90206208, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 519.2071249475703, "ttft_ms": 29.349292046390474, "decode_ms": 96.63545794319361, "receiver_prefill_ms": 0.0}, "total_wall_ms": 648.1490829028189, "ttft_ms": 29.349292046390474, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "848,090", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b5105383225daa63c1554bd08d37339a9b2d2a1066058fc15ab16cf75f05d898"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "6bd32ad68d33fdf71ed67375", "attempt_id": "de8fb4e4f82d", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:59:13.972338Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 90206208}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 53, "message_tokens": null, "message_utf8_bytes": 8736, "cache_tensor_bytes": 90206208, "serialized_payload_bytes": 90206208, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 519.2071249475703, "ttft_ms": 27.62079203967005, "decode_ms": 68.97866702638566, "receiver_prefill_ms": 0.0}, "total_wall_ms": 617.9721659282222, "ttft_ms": 27.62079203967005, "input_usage": {}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Brigam-1", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "cbeedb5efb33fd78c3c90ee9a3fd8829f55bdc158719dbd6bd04a8179d5b3aef"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "ef105a2850074c0228a546bb", "attempt_id": "9f2a531d27a4", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:59:14.086244Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 90206208}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 8736, "cache_tensor_bytes": 90206208, "serialized_payload_bytes": 90206208, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 519.2071249475703, "ttft_ms": 27.79291698243469, "decode_ms": 97.59066603146493, "receiver_prefill_ms": 0.0}, "total_wall_ms": 647.5464588729665, "ttft_ms": 27.79291698243469, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1524411", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "82c0515172df8dcdac00e953e3b0ba882e2f9aad7ecbc89997918d6b4f2ff1df"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "58f27368b8f43e59f700a3a1", "attempt_id": "d872b058c91f", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:59:14.229832Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 90206208}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 48, "message_tokens": null, "message_utf8_bytes": 8736, "cache_tensor_bytes": 90206208, "serialized_payload_bytes": 90206208, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 519.2071249475703, "ttft_ms": 28.592167072929442, "decode_ms": 13.696584035642445, "receiver_prefill_ms": 0.0}, "total_wall_ms": 564.6522090537474, "ttft_ms": 28.592167072929442, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c1659c591c127aed11ed272d", "attempt_id": "dc7e20c60d3a", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:59:14.290053Z", "task_family": "synthetic", "document_id": "syn_03003_2000", "document_hash": "dddca1a38e3224870fcfae927c701c400fd885c19123ea840ee06a76a140f407", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 90206208}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1866, "query_tokens": 49, "message_tokens": null, "message_utf8_bytes": 8736, "cache_tensor_bytes": 90206208, "serialized_payload_bytes": 90206208, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 519.2071249475703, "ttft_ms": 26.022500009275973, "decode_ms": 14.227917068637908, "receiver_prefill_ms": 0.0}, "total_wall_ms": 562.0963749242947, "ttft_ms": 26.022500009275973, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Gamtiszen-2", "bogus_amount": "848,090", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03003_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "e6cda9fcc720c375093ea063", "attempt_id": "96b2c05d8a35", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:59:15.268672Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 6322, "message_tokens": 6238, "message_utf8_bytes": 30518, "cache_tensor_bytes": null, "serialized_payload_bytes": 30518, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.26454206090420485, "ttft_ms": 13.689916930161417, "decode_ms": 136.2111249472946, "receiver_prefill_ms": 754.2456249939278}, "total_wall_ms": 914.0367500949651, "ttft_ms": 13.689916930161417, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "9,660,295 USD", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 0.6666666666666666}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ab6aeca92ccfafb7f81985bc441b5922c315c0d77d91478c2d754b5345b5a2b6", "output_hash": "9ac317774493c257d57e9ca8fa3afca757f9725dd762262435d6b268a92da7a0"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "0c5505bf376c82011c3eaa2b", "attempt_id": "ff00b05568f1", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:59:16.200468Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 6326, "message_tokens": 6238, "message_utf8_bytes": 30518, "cache_tensor_bytes": null, "serialized_payload_bytes": 30518, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.26454206090420485, "ttft_ms": 13.953500078059733, "decode_ms": 136.14241604227573, "receiver_prefill_ms": 751.5247500268742}, "total_wall_ms": 911.8759591365233, "ttft_ms": 13.953500078059733, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "NJ-38Y-8204", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ab6aeca92ccfafb7f81985bc441b5922c315c0d77d91478c2d754b5345b5a2b6", "output_hash": "ed38af85bd1c6dc6543c96eeae3d88aab9f40fd903082d29addfcf3dc185df3f"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b052d27e01b76ce1eb3fd580", "attempt_id": "872a029a7652", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:59:17.128855Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 6334, "message_tokens": 6238, "message_utf8_bytes": 30518, "cache_tensor_bytes": null, "serialized_payload_bytes": 30518, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.26454206090420485, "ttft_ms": 14.026790973730385, "decode_ms": 96.3186250301078, "receiver_prefill_ms": 777.3984579835087}, "total_wall_ms": 897.4857500288635, "ttft_ms": 14.026790973730385, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4646500", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ab6aeca92ccfafb7f81985bc441b5922c315c0d77d91478c2d754b5345b5a2b6", "output_hash": "22c7e5c0df41a8fcf90ee1e6ad966a7b73b0da3ca2ded8274a61b1221ad03f00"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "dc6e386bd140ebab743c4636", "attempt_id": "2b9f6c5cb287", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:59:18.044192Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 6326, "message_tokens": 6238, "message_utf8_bytes": 30518, "cache_tensor_bytes": null, "serialized_payload_bytes": 30518, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.26454206090420485, "ttft_ms": 13.90662498306483, "decode_ms": 124.27462497726083, "receiver_prefill_ms": 813.0579580320045}, "total_wall_ms": 961.2499170470983, "ttft_ms": 13.90662498306483, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1,642,260", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ab6aeca92ccfafb7f81985bc441b5922c315c0d77d91478c2d754b5345b5a2b6", "output_hash": "2bc4cbca0c91d790508324ed0e180fbe979a27c5991cdf8d1df4a0d6c6478128"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "24ec48dc518b4d9e5ef9e465", "attempt_id": "bd07050e1d6c", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:59:19.022899Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 6335, "message_tokens": 6238, "message_utf8_bytes": 30518, "cache_tensor_bytes": null, "serialized_payload_bytes": 30518, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.26454206090420485, "ttft_ms": 14.027833938598633, "decode_ms": 69.43604198750108, "receiver_prefill_ms": 837.7688750624657}, "total_wall_ms": 931.17262609303, "ttft_ms": 14.027833938598633, "input_usage": {}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Varard-12", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ab6aeca92ccfafb7f81985bc441b5922c315c0d77d91478c2d754b5345b5a2b6", "output_hash": "5b121a68ee6d66f04f20cbc0076c052393b985c2fa5041f6aadd2feb2925e691"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b2821ecaffc842fd7f88dabc", "attempt_id": "aea19b9a306c", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:59:19.973866Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 6327, "message_tokens": 6238, "message_utf8_bytes": 30518, "cache_tensor_bytes": null, "serialized_payload_bytes": 30518, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.26454206090420485, "ttft_ms": 14.024749980308115, "decode_ms": 110.40595907252282, "receiver_prefill_ms": 855.288999970071}, "total_wall_ms": 989.5893760258332, "ttft_ms": 14.024749980308115, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "14060295", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ab6aeca92ccfafb7f81985bc441b5922c315c0d77d91478c2d754b5345b5a2b6", "output_hash": "e1b19cd4363409df51da91ab06de45d1b3ea8ec038b6a6b72810b31f3716b7d0"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "13296ab1feee87cf823af412", "attempt_id": "fa4a76f106c7", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:59:20.981655Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 6327, "message_tokens": 6238, "message_utf8_bytes": 30518, "cache_tensor_bytes": null, "serialized_payload_bytes": 30518, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.26454206090420485, "ttft_ms": 13.930167071521282, "decode_ms": 14.252750086598098, "receiver_prefill_ms": 850.9299999568611}, "total_wall_ms": 888.9288341160864, "ttft_ms": 13.930167071521282, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ab6aeca92ccfafb7f81985bc441b5922c315c0d77d91478c2d754b5345b5a2b6", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "9dbabe4d9742c9f207a61fae", "attempt_id": "fbcebbf81878", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:59:21.888149Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 6329, "message_tokens": 6238, "message_utf8_bytes": 30518, "cache_tensor_bytes": null, "serialized_payload_bytes": 30518, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.26454206090420485, "ttft_ms": 14.201666926965117, "decode_ms": 14.050832949578762, "receiver_prefill_ms": 862.598416977562}, "total_wall_ms": 900.6020000670105, "ttft_ms": 14.201666926965117, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ab6aeca92ccfafb7f81985bc441b5922c315c0d77d91478c2d754b5345b5a2b6", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "0b7e907a33397bc72316686a", "attempt_id": "9c9148d09814", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:59:37.432995Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 1123, "message_tokens": 1037, "message_utf8_bytes": 2230, "cache_tensor_bytes": null, "serialized_payload_bytes": 2230, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14626.71737500932, "ttft_ms": 13.410083018243313, "decode_ms": 94.59533402696252, "receiver_prefill_ms": 121.83729198295623}, "total_wall_ms": 14858.395999995992, "ttft_ms": 13.410083018243313, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "9660295", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "9aac2601a07a9e7cd3490bd841f12bb968c9e24059f89e9482ce96f7d49bab33", "output_hash": "91988b1215586dd6899ef1d2a4168b0a54c0112d95ddff92bc07d79f5ac5ee86"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "fc628eb7d0adde99ffc7274d", "attempt_id": "466caf95f8a0", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:59:37.682541Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 1127, "message_tokens": 1037, "message_utf8_bytes": 2230, "cache_tensor_bytes": null, "serialized_payload_bytes": 2230, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14626.71737500932, "ttft_ms": 13.397500035353005, "decode_ms": 133.67704104166478, "receiver_prefill_ms": 121.60183303058147}, "total_wall_ms": 14897.127500036731, "ttft_ms": 13.397500035353005, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "NJ-38Y-8204", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "9aac2601a07a9e7cd3490bd841f12bb968c9e24059f89e9482ce96f7d49bab33", "output_hash": "ed38af85bd1c6dc6543c96eeae3d88aab9f40fd903082d29addfcf3dc185df3f"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "3ed52046a34bea96872a727b", "attempt_id": "36c5cd0678f8", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:59:37.971328Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 1135, "message_tokens": 1037, "message_utf8_bytes": 2230, "cache_tensor_bytes": null, "serialized_payload_bytes": 2230, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14626.71737500932, "ttft_ms": 13.826375012286007, "decode_ms": 93.77391694579273, "receiver_prefill_ms": 121.64533406030387}, "total_wall_ms": 14857.7030840097, "ttft_ms": 13.826375012286007, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4646500", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "9aac2601a07a9e7cd3490bd841f12bb968c9e24059f89e9482ce96f7d49bab33", "output_hash": "22c7e5c0df41a8fcf90ee1e6ad966a7b73b0da3ca2ded8274a61b1221ad03f00"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "38ac504a97f0d76c2b043f2f", "attempt_id": "94d5031da322", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:59:38.220157Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 1127, "message_tokens": 1037, "message_utf8_bytes": 2230, "cache_tensor_bytes": null, "serialized_payload_bytes": 2230, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14626.71737500932, "ttft_ms": 13.243334018625319, "decode_ms": 94.45350000169128, "receiver_prefill_ms": 120.78658293467015}, "total_wall_ms": 14856.970333028585, "ttft_ms": 13.243334018625319, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1642260", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "9aac2601a07a9e7cd3490bd841f12bb968c9e24059f89e9482ce96f7d49bab33", "output_hash": "08505337332e43191f2e55bf7664c5072c21fe3054a4238b7a4ab81901c3be71"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "fcae0f92a6ad39ec703e878d", "attempt_id": "ac4b1e5b2742", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:59:38.470830Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 1136, "message_tokens": 1037, "message_utf8_bytes": 2230, "cache_tensor_bytes": null, "serialized_payload_bytes": 2230, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14626.71737500932, "ttft_ms": 13.35395802743733, "decode_ms": 66.84445799328387, "receiver_prefill_ms": 121.58770801033825}, "total_wall_ms": 14830.256583052687, "ttft_ms": 13.35395802743733, "input_usage": {}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Varard-12", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "9aac2601a07a9e7cd3490bd841f12bb968c9e24059f89e9482ce96f7d49bab33", "output_hash": "5b121a68ee6d66f04f20cbc0076c052393b985c2fa5041f6aadd2feb2925e691"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c25719c14800ee1ec29f9ce7", "attempt_id": "5b10bdefe151", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:59:38.692648Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 1128, "message_tokens": 1037, "message_utf8_bytes": 2230, "cache_tensor_bytes": null, "serialized_payload_bytes": 2230, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14626.71737500932, "ttft_ms": 13.699667062610388, "decode_ms": 93.4889999916777, "receiver_prefill_ms": 123.38908389210701}, "total_wall_ms": 14859.075916116126, "ttft_ms": 13.699667062610388, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1106155", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "9aac2601a07a9e7cd3490bd841f12bb968c9e24059f89e9482ce96f7d49bab33", "output_hash": "b55e05eac4185e76df0edf3adc8c910a7510bbab0878784ad6f979ce60ce13e2"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "4e86ee96ff28064eef784029", "attempt_id": "3468cd06ecc7", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:59:38.945721Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 1128, "message_tokens": 1037, "message_utf8_bytes": 2230, "cache_tensor_bytes": null, "serialized_payload_bytes": 2230, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14626.71737500932, "ttft_ms": 13.411334017291665, "decode_ms": 13.67258292157203, "receiver_prefill_ms": 131.05583400465548}, "total_wall_ms": 14786.587458103895, "ttft_ms": 13.411334017291665, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "9aac2601a07a9e7cd3490bd841f12bb968c9e24059f89e9482ce96f7d49bab33", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "9c0221c64f52180dbbad46ae", "attempt_id": "5f653e0e9c60", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:59:39.123442Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 1130, "message_tokens": 1037, "message_utf8_bytes": 2230, "cache_tensor_bytes": null, "serialized_payload_bytes": 2230, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14626.71737500932, "ttft_ms": 13.587292050942779, "decode_ms": 13.466790900565684, "receiver_prefill_ms": 136.99754199478775}, "total_wall_ms": 14792.52183297649, "ttft_ms": 13.587292050942779, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "9aac2601a07a9e7cd3490bd841f12bb968c9e24059f89e9482ce96f7d49bab33", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "827b5572fab260689c911d55", "attempt_id": "883e107a9dc2", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:59:54.112589Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 13}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 623, "message_tokens": 534, "message_utf8_bytes": 1404, "cache_tensor_bytes": null, "serialized_payload_bytes": 1404, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14803.205291973427, "ttft_ms": 13.332791975699365, "decode_ms": 118.44683298841119, "receiver_prefill_ms": 73.71050002984703}, "total_wall_ms": 15010.005875024945, "ttft_ms": 13.332791975699365, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1,449,850", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "1a1e262b1177f03c2e9a63280ede99d3433c1f04d114dbe32df04ea048c9a244", "output_hash": "ebc73cc2db51bb11b57ba154709857a98a2e20df8b51e7ff500d8c7268c7a05e"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "6a930321119aa940fd1eb29e", "attempt_id": "2c71e51c2857", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:59:54.337759Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 13}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 627, "message_tokens": 534, "message_utf8_bytes": 1404, "cache_tensor_bytes": null, "serialized_payload_bytes": 1404, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14803.205291973427, "ttft_ms": 13.05433304514736, "decode_ms": 13.57554190326482, "receiver_prefill_ms": 73.60204204451293}, "total_wall_ms": 14904.705416993238, "ttft_ms": 13.05433304514736, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "1a1e262b1177f03c2e9a63280ede99d3433c1f04d114dbe32df04ea048c9a244", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "bb6ec34cd9353e5096244259", "attempt_id": "3c9875a62fc1", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:59:54.456894Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 13}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 635, "message_tokens": 534, "message_utf8_bytes": 1404, "cache_tensor_bytes": null, "serialized_payload_bytes": 1404, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14803.205291973427, "ttft_ms": 13.486332958564162, "decode_ms": 92.62945898808539, "receiver_prefill_ms": 73.49037507083267}, "total_wall_ms": 14984.032501000911, "ttft_ms": 13.486332958564162, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4896700", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "1a1e262b1177f03c2e9a63280ede99d3433c1f04d114dbe32df04ea048c9a244", "output_hash": "7b473942a50c895967c7191b655a6f5a08b11a0e190b4f9f586ae6c445ebffc4"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "121cff48df7a4f91b72eb005", "attempt_id": "48b8a920df87", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:59:54.655538Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 13}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 627, "message_tokens": 534, "message_utf8_bytes": 1404, "cache_tensor_bytes": null, "serialized_payload_bytes": 1404, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14803.205291973427, "ttft_ms": 13.146833050996065, "decode_ms": 118.68345900438726, "receiver_prefill_ms": 72.93904200196266}, "total_wall_ms": 15009.251000010408, "ttft_ms": 13.146833050996065, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1,389,935", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "1a1e262b1177f03c2e9a63280ede99d3433c1f04d114dbe32df04ea048c9a244", "output_hash": "ede95cf26de2e7b5fed721efb7107c4c59e9471b34a4d3fd4451b424fcd8d73f"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "480c44bd1639f54c6f3ea715", "attempt_id": "c1182e4b9723", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:59:54.879234Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 13}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 636, "message_tokens": 534, "message_utf8_bytes": 1404, "cache_tensor_bytes": null, "serialized_payload_bytes": 1404, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14803.205291973427, "ttft_ms": 13.43187503516674, "decode_ms": 91.74337494187057, "receiver_prefill_ms": 72.9988330276683}, "total_wall_ms": 14982.653209008276, "ttft_ms": 13.43187503516674, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Obysellith-16", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "1a1e262b1177f03c2e9a63280ede99d3433c1f04d114dbe32df04ea048c9a244", "output_hash": "66b39064cfe9c103d8574722e2dac0e9a19ced28c716ed2067d84dd8796e7762"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a6df3109a257935445f8e5b8", "attempt_id": "81beea1055b5", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:59:55.076459Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 13}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 628, "message_tokens": 534, "message_utf8_bytes": 1404, "cache_tensor_bytes": null, "serialized_payload_bytes": 1404, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14803.205291973427, "ttft_ms": 13.263457920402288, "decode_ms": 92.21254207659513, "receiver_prefill_ms": 73.01900000311434}, "total_wall_ms": 14982.949124998413, "ttft_ms": 13.263457920402288, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1155845", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "1a1e262b1177f03c2e9a63280ede99d3433c1f04d114dbe32df04ea048c9a244", "output_hash": "309acb245bdb0fbc49045d9c510553c6e54bfe28796ed00a3be25d748275892b"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "55a1dc0caca843e9c0f2aba0", "attempt_id": "4348a1333d03", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:59:55.276571Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 13}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 628, "message_tokens": 534, "message_utf8_bytes": 1404, "cache_tensor_bytes": null, "serialized_payload_bytes": 1404, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14803.205291973427, "ttft_ms": 13.27358407434076, "decode_ms": 13.021750026382506, "receiver_prefill_ms": 73.07129097171128}, "total_wall_ms": 14903.802708955482, "ttft_ms": 13.27358407434076, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "1a1e262b1177f03c2e9a63280ede99d3433c1f04d114dbe32df04ea048c9a244", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b46a4860f222f78d8467aa76", "attempt_id": "96783645748d", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T07:59:55.395334Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 13}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 630, "message_tokens": 534, "message_utf8_bytes": 1404, "cache_tensor_bytes": null, "serialized_payload_bytes": 1404, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14803.205291973427, "ttft_ms": 13.078083051368594, "decode_ms": 13.164667063392699, "receiver_prefill_ms": 73.05608305614442}, "total_wall_ms": 14903.751083882526, "ttft_ms": 13.078083051368594, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "1a1e262b1177f03c2e9a63280ede99d3433c1f04d114dbe32df04ea048c9a244", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "8bd7caf84b3a58a098cbdeff", "attempt_id": "2b6b4bcdf656", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:00:10.123089Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 1413, "message_tokens": 1325, "message_utf8_bytes": 3347, "cache_tensor_bytes": null, "serialized_payload_bytes": 3347, "side_information_bytes": 0, "fetched_bytes": 1863, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14609.214707976207, "ttft_ms": 13.374082976952195, "decode_ms": 93.69945898652077, "receiver_prefill_ms": 151.66870795655996}, "total_wall_ms": 14870.103750028647, "ttft_ms": 13.374082976952195, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "9660295", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "d23627eaddf2b4e204c59b642d09524c24b88423ef84e25bde487e84d73fc9ec", "output_hash": "91988b1215586dd6899ef1d2a4168b0a54c0112d95ddff92bc07d79f5ac5ee86"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "1df883d679b3900e63fcc281", "attempt_id": "a008fb3be01d", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:00:10.404883Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 1365, "message_tokens": 1273, "message_utf8_bytes": 3262, "cache_tensor_bytes": null, "serialized_payload_bytes": 3262, "side_information_bytes": 0, "fetched_bytes": 1783, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14609.214707976207, "ttft_ms": 13.420625007711351, "decode_ms": 13.299082987941802, "receiver_prefill_ms": 144.72279197070748}, "total_wall_ms": 14782.742333016358, "ttft_ms": 13.420625007711351, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "f365f53bdd4c5101370c17dba068c1de6fa13a62ab163c4cea0f5f53e6cf987d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "fea8e2506a26fafcd4467175", "attempt_id": "dd195ee6d3b1", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:00:10.596091Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 1364, "message_tokens": 1264, "message_utf8_bytes": 3237, "cache_tensor_bytes": null, "serialized_payload_bytes": 3237, "side_information_bytes": 0, "fetched_bytes": 1757, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14609.214707976207, "ttft_ms": 13.19783297367394, "decode_ms": 93.34908402524889, "receiver_prefill_ms": 144.35837499331683}, "total_wall_ms": 14862.266332958825, "ttft_ms": 13.19783297367394, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4646500", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "1c100ea84735778b1fc18d4d490aff56833352a25ab82c40177661bc123e257d", "output_hash": "22c7e5c0df41a8fcf90ee1e6ad966a7b73b0da3ca2ded8274a61b1221ad03f00"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b886de4c319ca4b2f3820592", "attempt_id": "125a5057c5ba", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:00:10.866472Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 1365, "message_tokens": 1273, "message_utf8_bytes": 3262, "cache_tensor_bytes": null, "serialized_payload_bytes": 3262, "side_information_bytes": 0, "fetched_bytes": 1783, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14609.214707976207, "ttft_ms": 13.568292022682726, "decode_ms": 94.18683394324034, "receiver_prefill_ms": 150.1662499504164}, "total_wall_ms": 14869.25470805727, "ttft_ms": 13.568292022682726, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1642260", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "09985158617adce75a537d21d470213451a47ef4cade95a567fd8383ac614dc1", "output_hash": "08505337332e43191f2e55bf7664c5072c21fe3054a4238b7a4ab81901c3be71"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "cf2c6109ef387c6cd85faa5c", "attempt_id": "b7cbade79737", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:00:11.143989Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 1396, "message_tokens": 1295, "message_utf8_bytes": 3299, "cache_tensor_bytes": null, "serialized_payload_bytes": 3299, "side_information_bytes": 0, "fetched_bytes": 1821, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14609.214707976207, "ttft_ms": 13.61000002361834, "decode_ms": 93.74258294701576, "receiver_prefill_ms": 165.41758307721466}, "total_wall_ms": 14884.04695794452, "ttft_ms": 13.61000002361834, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Obysellith-16", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "592f4202e8037247d8f010ff78ec93cb8fc6ecb05adc2e0f4e276996bb0934e5", "output_hash": "66b39064cfe9c103d8574722e2dac0e9a19ced28c716ed2067d84dd8796e7762"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "3256e34898999fa02b2c76b0", "attempt_id": "8fbb90358f78", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:00:11.439096Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 1427, "message_tokens": 1334, "message_utf8_bytes": 3343, "cache_tensor_bytes": null, "serialized_payload_bytes": 3343, "side_information_bytes": 0, "fetched_bytes": 1860, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14609.214707976207, "ttft_ms": 13.623708044178784, "decode_ms": 107.73258400149643, "receiver_prefill_ms": 170.6397089874372}, "total_wall_ms": 14903.355167014524, "ttft_ms": 13.623708044178784, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "10648015", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "f2fdd31aca64e10210ee81a046cf3511a7b8e81d19d84bc99f5301029951a634", "output_hash": "69cf70247e72c000a18a32db8c4ce9c22838070d500f56e45a83572bc5e8584b"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "2387d7eb57ef01bd00e6c81a", "attempt_id": "fcaa950cff3b", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:00:11.751066Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 1366, "message_tokens": 1273, "message_utf8_bytes": 3262, "cache_tensor_bytes": null, "serialized_payload_bytes": 3262, "side_information_bytes": 0, "fetched_bytes": 1783, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14609.214707976207, "ttft_ms": 13.575208955444396, "decode_ms": 13.437333982437849, "receiver_prefill_ms": 160.85649991873652}, "total_wall_ms": 14799.127873964608, "ttft_ms": 13.575208955444396, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "f365f53bdd4c5101370c17dba068c1de6fa13a62ab163c4cea0f5f53e6cf987d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "0697778d71a0b0a7f6243a0d", "attempt_id": "a599f4f62416", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:00:11.958800Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 1368, "message_tokens": 1273, "message_utf8_bytes": 3262, "cache_tensor_bytes": null, "serialized_payload_bytes": 3262, "side_information_bytes": 0, "fetched_bytes": 1783, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14609.214707976207, "ttft_ms": 13.81687493994832, "decode_ms": 13.37708393111825, "receiver_prefill_ms": 168.1108339689672}, "total_wall_ms": 14806.537623982877, "ttft_ms": 13.81687493994832, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "f365f53bdd4c5101370c17dba068c1de6fa13a62ab163c4cea0f5f53e6cf987d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "2218dd041631453c425be0f6", "attempt_id": "561b6a405921", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:00:13.087964Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 232833024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 42, "message_tokens": null, "message_utf8_bytes": 30492, "cache_tensor_bytes": 232833024, "serialized_payload_bytes": 232833024, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 914.2973750131205, "ttft_ms": 53.529666969552636, "decode_ms": 135.801124968566, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1113.3928339695558, "ttft_ms": 53.529666969552636, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "9,660,295", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "bda34df0df73c55ae84212f4401dad18f66af356a6de893b4a440316d26ecdfa"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c44b427217958530b4b1821e", "attempt_id": "c6dc6c4dcfb3", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:00:13.307619Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 232833024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 46, "message_tokens": null, "message_utf8_bytes": 30492, "cache_tensor_bytes": 232833024, "serialized_payload_bytes": 232833024, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 914.2973750131205, "ttft_ms": 49.44629198871553, "decode_ms": 165.06183298770338, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1136.3888749619946, "ttft_ms": 49.44629198871553, "input_usage": {}, "output_usage": {"output_tokens": 11}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "YF-41T-2802", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "2c942bb83ffbc6ef73a6f94c0666c0c3cc5be2520af627409601e33b0571552b"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "aa6df15152a0c547b2a4b784", "attempt_id": "6626893a36ab", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:00:13.547181Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 232833024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 54, "message_tokens": null, "message_utf8_bytes": 30492, "cache_tensor_bytes": 232833024, "serialized_payload_bytes": 232833024, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 914.2973750131205, "ttft_ms": 48.06212510447949, "decode_ms": 104.29058293811977, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1074.2298749973997, "ttft_ms": 48.06212510447949, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7067400", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "7f470ea83ee0eb2427026d799503d40b32454ebe1ef64f12c8d1a4a53830a524"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "f9efd2eb524de976bc7820bc", "attempt_id": "6a34a92611f0", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:00:13.724825Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 232833024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 46, "message_tokens": null, "message_utf8_bytes": 30492, "cache_tensor_bytes": 232833024, "serialized_payload_bytes": 232833024, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 914.2973750131205, "ttft_ms": 48.619500012136996, "decode_ms": 134.6581670222804, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1104.1922500589862, "ttft_ms": 48.619500012136996, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1,642,260", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "2bc4cbca0c91d790508324ed0e180fbe979a27c5991cdf8d1df4a0d6c6478128"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "e3002c8ef28b01be05ef09d1", "attempt_id": "36538d5485e8", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:00:13.932446Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 232833024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 55, "message_tokens": null, "message_utf8_bytes": 30492, "cache_tensor_bytes": 232833024, "serialized_payload_bytes": 232833024, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 914.2973750131205, "ttft_ms": 48.08874998707324, "decode_ms": 104.34162500314415, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1074.3226250633597, "ttft_ms": 48.08874998707324, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Obysellith-16", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "66b39064cfe9c103d8574722e2dac0e9a19ced28c716ed2067d84dd8796e7762"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "3941391d986a566590a9289d", "attempt_id": "0cb650b0c0a9", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:00:14.110226Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 232833024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 30492, "cache_tensor_bytes": 232833024, "serialized_payload_bytes": 232833024, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 914.2973750131205, "ttft_ms": 48.60233305953443, "decode_ms": 105.05658306647092, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1075.5586669547483, "ttft_ms": 48.60233305953443, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "3369169", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "ce970ae9f35f6e4f8e44b99e0cd5793aa6668fad23c4ea246e78fbb6d70011ed"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "dfaefd9565aa43ac9343782e", "attempt_id": "ad70ad65c38b", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:00:14.289331Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 232833024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 30492, "cache_tensor_bytes": 232833024, "serialized_payload_bytes": 232833024, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 914.2973750131205, "ttft_ms": 48.40712493751198, "decode_ms": 15.087292063981295, "receiver_prefill_ms": 0.0}, "total_wall_ms": 985.4115409543738, "ttft_ms": 48.40712493751198, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "0164c42a7722caca10dd7968", "attempt_id": "9e32acd490df", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:00:14.377761Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 232833024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 49, "message_tokens": null, "message_utf8_bytes": 30492, "cache_tensor_bytes": 232833024, "serialized_payload_bytes": 232833024, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 914.2973750131205, "ttft_ms": 49.332624999806285, "decode_ms": 14.856499969027936, "receiver_prefill_ms": 0.0}, "total_wall_ms": 986.1301670316607, "ttft_ms": 49.332624999806285, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "9eb1096334c0ce6b00c1025e", "attempt_id": "46c93bb787a6", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:00:14.468104Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 6436, "message_tokens": 6352, "message_utf8_bytes": 31022, "cache_tensor_bytes": null, "serialized_payload_bytes": 31022, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.028416980057954788, "ttft_ms": 14.23379103653133, "decode_ms": 96.27420792821795, "receiver_prefill_ms": 784.8512079799548}, "total_wall_ms": 905.1070420537144, "ttft_ms": 14.23379103653133, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "9660295", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ae650e04bf513cf0585b912ff9050b48e47afad12e291aa24fefbf1fd19617f1", "output_hash": "91988b1215586dd6899ef1d2a4168b0a54c0112d95ddff92bc07d79f5ac5ee86"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a74ff7b572a8e707ae588910", "attempt_id": "47b594b51c08", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:00:15.390363Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 6440, "message_tokens": 6352, "message_utf8_bytes": 31022, "cache_tensor_bytes": null, "serialized_payload_bytes": 31022, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.028416980057954788, "ttft_ms": 14.095625025220215, "decode_ms": 138.1268339464441, "receiver_prefill_ms": 816.6690419893712}, "total_wall_ms": 978.7212079390883, "ttft_ms": 14.095625025220215, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "NJ-38Y-8204", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ae650e04bf513cf0585b912ff9050b48e47afad12e291aa24fefbf1fd19617f1", "output_hash": "ed38af85bd1c6dc6543c96eeae3d88aab9f40fd903082d29addfcf3dc185df3f"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "692effd39b8fce085cb28e68", "attempt_id": "9acc90c99bf0", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:00:16.386461Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 6448, "message_tokens": 6352, "message_utf8_bytes": 31022, "cache_tensor_bytes": null, "serialized_payload_bytes": 31022, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.028416980057954788, "ttft_ms": 13.909582979977131, "decode_ms": 97.80454100109637, "receiver_prefill_ms": 841.5965420426801}, "total_wall_ms": 962.9149590618908, "ttft_ms": 13.909582979977131, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4646500", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ae650e04bf513cf0585b912ff9050b48e47afad12e291aa24fefbf1fd19617f1", "output_hash": "22c7e5c0df41a8fcf90ee1e6ad966a7b73b0da3ca2ded8274a61b1221ad03f00"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "31ebc154ef73aefb29742d0f", "attempt_id": "1fd3798a7c9f", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:00:17.366803Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 6440, "message_tokens": 6352, "message_utf8_bytes": 31022, "cache_tensor_bytes": null, "serialized_payload_bytes": 31022, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.028416980057954788, "ttft_ms": 13.982499949634075, "decode_ms": 125.72762498166412, "receiver_prefill_ms": 865.0843750219792}, "total_wall_ms": 1014.6056669764221, "ttft_ms": 13.982499949634075, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1,642,260", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ae650e04bf513cf0585b912ff9050b48e47afad12e291aa24fefbf1fd19617f1", "output_hash": "2bc4cbca0c91d790508324ed0e180fbe979a27c5991cdf8d1df4a0d6c6478128"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "294c56c65af1f178da8f4e7d", "attempt_id": "7b7e34eb3327", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:00:18.399073Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 6449, "message_tokens": 6352, "message_utf8_bytes": 31022, "cache_tensor_bytes": null, "serialized_payload_bytes": 31022, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.028416980057954788, "ttft_ms": 14.244875055737793, "decode_ms": 69.66999999713153, "receiver_prefill_ms": 862.1578339952976}, "total_wall_ms": 955.7945419801399, "ttft_ms": 14.244875055737793, "input_usage": {}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Varard-12", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ae650e04bf513cf0585b912ff9050b48e47afad12e291aa24fefbf1fd19617f1", "output_hash": "5b121a68ee6d66f04f20cbc0076c052393b985c2fa5041f6aadd2feb2925e691"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "1671a5ac0391c3590e4d22e5", "attempt_id": "ade81f2c6550", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:00:19.372386Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 6441, "message_tokens": 6352, "message_utf8_bytes": 31022, "cache_tensor_bytes": null, "serialized_payload_bytes": 31022, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.028416980057954788, "ttft_ms": 13.95874994341284, "decode_ms": 110.30370800290257, "receiver_prefill_ms": 890.6472501112148}, "total_wall_ms": 1024.6112500317395, "ttft_ms": 13.95874994341284, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "12762.67", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ae650e04bf513cf0585b912ff9050b48e47afad12e291aa24fefbf1fd19617f1", "output_hash": "db80f2f15d821b39466eede3e153e920f82cea8c2ca5e38806f0cd1fbe638697"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "31b2daa40b84d1c67873a378", "attempt_id": "6488fb8178da", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:00:20.414471Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 6441, "message_tokens": 6352, "message_utf8_bytes": 31022, "cache_tensor_bytes": null, "serialized_payload_bytes": 31022, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.028416980057954788, "ttft_ms": 13.971750042401254, "decode_ms": 14.126833993941545, "receiver_prefill_ms": 854.8777080141008}, "total_wall_ms": 892.6099588861689, "ttft_ms": 13.971750042401254, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ae650e04bf513cf0585b912ff9050b48e47afad12e291aa24fefbf1fd19617f1", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "8dc361b249788ecaab4feb2f", "attempt_id": "08975443c4ca", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:00:21.324412Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 6443, "message_tokens": 6352, "message_utf8_bytes": 31022, "cache_tensor_bytes": null, "serialized_payload_bytes": 31022, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.028416980057954788, "ttft_ms": 13.983834069222212, "decode_ms": 14.105374924838543, "receiver_prefill_ms": 877.5219590170309}, "total_wall_ms": 915.3205839684233, "ttft_ms": 13.983834069222212, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ae650e04bf513cf0585b912ff9050b48e47afad12e291aa24fefbf1fd19617f1", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "e6e011401838349d22034385", "attempt_id": "651047ba3681", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:00:47.187786Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1801}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 1911, "message_tokens": 1826, "message_utf8_bytes": 5945, "cache_tensor_bytes": null, "serialized_payload_bytes": 5945, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24930.066749919206, "ttft_ms": 13.422041898593307, "decode_ms": 93.7512080417946, "receiver_prefill_ms": 201.89337502233684}, "total_wall_ms": 25247.038791887462, "ttft_ms": 13.422041898593307, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "9660295", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "c07066c6a379df127af044fa761edde13a8ee0c6380e1bb0d9712f5fe44a0efe", "output_hash": "91988b1215586dd6899ef1d2a4168b0a54c0112d95ddff92bc07d79f5ac5ee86"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b86643cdb6164899f8514e9f", "attempt_id": "0c332ac509f4", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:00:47.523400Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1801}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 1915, "message_tokens": 1826, "message_utf8_bytes": 5945, "cache_tensor_bytes": null, "serialized_payload_bytes": 5945, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24930.066749919206, "ttft_ms": 13.922165962867439, "decode_ms": 133.2146660424769, "receiver_prefill_ms": 202.2248749854043}, "total_wall_ms": 25282.298249891028, "ttft_ms": 13.922165962867439, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "NJ-38Y-8204", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "c07066c6a379df127af044fa761edde13a8ee0c6380e1bb0d9712f5fe44a0efe", "output_hash": "ed38af85bd1c6dc6543c96eeae3d88aab9f40fd903082d29addfcf3dc185df3f"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "218084e1e0482711e86f5b2f", "attempt_id": "43b736715d53", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:00:47.893425Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1801}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 1923, "message_tokens": 1826, "message_utf8_bytes": 5945, "cache_tensor_bytes": null, "serialized_payload_bytes": 5945, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24930.066749919206, "ttft_ms": 13.81537492852658, "decode_ms": 94.53795896843076, "receiver_prefill_ms": 215.56795900687575}, "total_wall_ms": 25256.805915851146, "ttft_ms": 13.81537492852658, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4646500", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "c07066c6a379df127af044fa761edde13a8ee0c6380e1bb0d9712f5fe44a0efe", "output_hash": "22c7e5c0df41a8fcf90ee1e6ad966a7b73b0da3ca2ded8274a61b1221ad03f00"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "802396c47339fadec9b74647", "attempt_id": "2d0c55a8cba4", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:00:48.238179Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1801}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 1915, "message_tokens": 1826, "message_utf8_bytes": 5945, "cache_tensor_bytes": null, "serialized_payload_bytes": 5945, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24930.066749919206, "ttft_ms": 13.856083038263023, "decode_ms": 121.59237498417497, "receiver_prefill_ms": 236.88445799052715}, "total_wall_ms": 25305.233707884327, "ttft_ms": 13.856083038263023, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1,642,260", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "c07066c6a379df127af044fa761edde13a8ee0c6380e1bb0d9712f5fe44a0efe", "output_hash": "2bc4cbca0c91d790508324ed0e180fbe979a27c5991cdf8d1df4a0d6c6478128"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "aee8d1ef0311f8e5b4f1cfd6", "attempt_id": "22c60ab237da", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:00:48.631344Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1801}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 1924, "message_tokens": 1826, "message_utf8_bytes": 5945, "cache_tensor_bytes": null, "serialized_payload_bytes": 5945, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24930.066749919206, "ttft_ms": 13.575625023804605, "decode_ms": 68.02124995738268, "receiver_prefill_ms": 233.02054195664823}, "total_wall_ms": 25247.593209031038, "ttft_ms": 13.575625023804605, "input_usage": {}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Varard-12", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "c07066c6a379df127af044fa761edde13a8ee0c6380e1bb0d9712f5fe44a0efe", "output_hash": "5b121a68ee6d66f04f20cbc0076c052393b985c2fa5041f6aadd2feb2925e691"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "92d575e7455115c81dbb3e88", "attempt_id": "10a5357b0d10", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:00:48.966536Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1801}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 1916, "message_tokens": 1826, "message_utf8_bytes": 5945, "cache_tensor_bytes": null, "serialized_payload_bytes": 5945, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24930.066749919206, "ttft_ms": 13.875082950107753, "decode_ms": 95.13354196678847, "receiver_prefill_ms": 234.15108397603035}, "total_wall_ms": 25276.1176249478, "ttft_ms": 13.875082950107753, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5409371", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "c07066c6a379df127af044fa761edde13a8ee0c6380e1bb0d9712f5fe44a0efe", "output_hash": "bdca170867f717893400924d414afbd8f170448483bb18861583e48db9d59933"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "5f638f4105a8ad74dc58d3cd", "attempt_id": "f6c5f0e7f485", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:00:49.330819Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1801}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 1916, "message_tokens": 1826, "message_utf8_bytes": 5945, "cache_tensor_bytes": null, "serialized_payload_bytes": 5945, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24930.066749919206, "ttft_ms": 13.686708989553154, "decode_ms": 13.463458977639675, "receiver_prefill_ms": 229.56129198428243}, "total_wall_ms": 25189.6107919747, "ttft_ms": 13.686708989553154, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "c07066c6a379df127af044fa761edde13a8ee0c6380e1bb0d9712f5fe44a0efe", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "4de3ce50c1e72f9add34c471", "attempt_id": "3eb3df20164e", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:00:49.608337Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1801}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 1918, "message_tokens": 1826, "message_utf8_bytes": 5945, "cache_tensor_bytes": null, "serialized_payload_bytes": 5945, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24930.066749919206, "ttft_ms": 13.965500053018332, "decode_ms": 14.025541953742504, "receiver_prefill_ms": 249.69879095442593}, "total_wall_ms": 25210.607583867386, "ttft_ms": 13.965500053018332, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "c07066c6a379df127af044fa761edde13a8ee0c6380e1bb0d9712f5fe44a0efe", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "22996ec904adf734106bb90a", "attempt_id": "476c421f6eb6", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:01:15.686112Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 27}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 1159, "message_tokens": 1070, "message_utf8_bytes": 2812, "cache_tensor_bytes": null, "serialized_payload_bytes": 2812, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25779.04012496583, "ttft_ms": 13.607457978650928, "decode_ms": 13.373875059187412, "receiver_prefill_ms": 531.2299579381943}, "total_wall_ms": 26339.08620791044, "ttft_ms": 13.607457978650928, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "6fa1438434f1b45af1e5831de53beffdecf19ae306475aa76bab613e3237fe39", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "153a28c94465d8ce776bca3d", "attempt_id": "f8943de80aaa", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:01:16.268640Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 27}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 1163, "message_tokens": 1070, "message_utf8_bytes": 2812, "cache_tensor_bytes": null, "serialized_payload_bytes": 2812, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25779.04012496583, "ttft_ms": 35.25166597682983, "decode_ms": 13.719500042498112, "receiver_prefill_ms": 1220.6426250049844}, "total_wall_ms": 27050.6173329195, "ttft_ms": 35.25166597682983, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "6fa1438434f1b45af1e5831de53beffdecf19ae306475aa76bab613e3237fe39", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "ffe83b9887e91a080b6af416", "attempt_id": "461cd98d75e6", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:01:17.559173Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 27}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 1171, "message_tokens": 1070, "message_utf8_bytes": 2812, "cache_tensor_bytes": null, "serialized_payload_bytes": 2812, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25779.04012496583, "ttft_ms": 13.709708000533283, "decode_ms": 97.6259580347687, "receiver_prefill_ms": 141.36116695590317}, "total_wall_ms": 26033.72520802077, "ttft_ms": 13.709708000533283, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4646500", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "6fa1438434f1b45af1e5831de53beffdecf19ae306475aa76bab613e3237fe39", "output_hash": "22c7e5c0df41a8fcf90ee1e6ad966a7b73b0da3ca2ded8274a61b1221ad03f00"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "1932491bb7bca49f581d8bee", "attempt_id": "3b5b3cb2e9e3", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:01:17.838967Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 27}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 1163, "message_tokens": 1070, "message_utf8_bytes": 2812, "cache_tensor_bytes": null, "serialized_payload_bytes": 2812, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25779.04012496583, "ttft_ms": 13.810041942633688, "decode_ms": 120.87525008246303, "receiver_prefill_ms": 124.90675004664809}, "total_wall_ms": 26040.479208924808, "ttft_ms": 13.810041942633688, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1,389,935", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "6fa1438434f1b45af1e5831de53beffdecf19ae306475aa76bab613e3237fe39", "output_hash": "ede95cf26de2e7b5fed721efb7107c4c59e9471b34a4d3fd4451b424fcd8d73f"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "1d7cb3702457a4f1fe460c6e", "attempt_id": "bb4b29a085b3", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:01:18.122893Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 27}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 1172, "message_tokens": 1070, "message_utf8_bytes": 2812, "cache_tensor_bytes": null, "serialized_payload_bytes": 2812, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25779.04012496583, "ttft_ms": 13.292707968503237, "decode_ms": 92.46633399743587, "receiver_prefill_ms": 124.78349998127669}, "total_wall_ms": 26011.415833956562, "ttft_ms": 13.292707968503237, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Obysellith-16", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "6fa1438434f1b45af1e5831de53beffdecf19ae306475aa76bab613e3237fe39", "output_hash": "66b39064cfe9c103d8574722e2dac0e9a19ced28c716ed2067d84dd8796e7762"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "e60d42cb01e806cb2db4bb36", "attempt_id": "0b0149d71a2a", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:01:18.373611Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 27}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 1164, "message_tokens": 1070, "message_utf8_bytes": 2812, "cache_tensor_bytes": null, "serialized_payload_bytes": 2812, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25779.04012496583, "ttft_ms": 13.286749948747456, "decode_ms": 92.58666599635035, "receiver_prefill_ms": 124.48970798868686}, "total_wall_ms": 26011.147083947435, "ttft_ms": 13.286749948747456, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7589171", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "6fa1438434f1b45af1e5831de53beffdecf19ae306475aa76bab613e3237fe39", "output_hash": "5f033e151985c1be5197f76714232888aefc79debfde9e069e241225f84fae78"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "660e14557b906e840af185d6", "attempt_id": "19de158203ca", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:01:18.623407Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 27}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 1164, "message_tokens": 1070, "message_utf8_bytes": 2812, "cache_tensor_bytes": null, "serialized_payload_bytes": 2812, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25779.04012496583, "ttft_ms": 13.153833104297519, "decode_ms": 13.26558308210224, "receiver_prefill_ms": 124.22904092818499}, "total_wall_ms": 25931.439250009134, "ttft_ms": 13.153833104297519, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "6fa1438434f1b45af1e5831de53beffdecf19ae306475aa76bab613e3237fe39", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "4f132f3cf22a4fe3ac6b1adb", "attempt_id": "5942ad8d00f0", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:01:18.794231Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 27}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 1166, "message_tokens": 1070, "message_utf8_bytes": 2812, "cache_tensor_bytes": null, "serialized_payload_bytes": 2812, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25779.04012496583, "ttft_ms": 14.03187494724989, "decode_ms": 14.064458897337317, "receiver_prefill_ms": 154.04266689438373}, "total_wall_ms": 25962.99879089929, "ttft_ms": 14.03187494724989, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "6fa1438434f1b45af1e5831de53beffdecf19ae306475aa76bab613e3237fe39", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "f1a6e712d358579226b6425c", "attempt_id": "b091fb54e9af", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:01:45.026323Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 2554, "message_tokens": 2466, "message_utf8_bytes": 6335, "cache_tensor_bytes": null, "serialized_payload_bytes": 6335, "side_information_bytes": 0, "fetched_bytes": 3453, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26030.547000002116, "ttft_ms": 13.335582916624844, "decode_ms": 94.88445799797773, "receiver_prefill_ms": 279.7011659713462}, "total_wall_ms": 26422.41537501104, "ttft_ms": 13.335582916624844, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "9660295", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ba5bd2d87a4acba33bb8d8917a69a121053bd1ef211fed01a67829f83852f500", "output_hash": "91988b1215586dd6899ef1d2a4168b0a54c0112d95ddff92bc07d79f5ac5ee86"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "328bebd56b9a4f20dbabf54d", "attempt_id": "face7b567a65", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:01:45.441035Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 2507, "message_tokens": 2415, "message_utf8_bytes": 6199, "cache_tensor_bytes": null, "serialized_payload_bytes": 6199, "side_information_bytes": 0, "fetched_bytes": 3313, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26030.547000002116, "ttft_ms": 13.92774994019419, "decode_ms": 137.3677080264315, "receiver_prefill_ms": 286.6094590863213}, "total_wall_ms": 26471.686042030342, "ttft_ms": 13.92774994019419, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "TS-42C-2098", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ad6a27b48743ac5b8905391d9b2049e9ecb55e87b8199c1c9e9ae0a7a4ae7faf", "output_hash": "51e8fd0b8140b0749334ad9ccec8a67ee7be75dd76dada04f3ad033bf9b9b10f"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b55f93837d91ccfe0e67062f", "attempt_id": "f3134889b201", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:01:45.900611Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 2566, "message_tokens": 2466, "message_utf8_bytes": 6314, "cache_tensor_bytes": null, "serialized_payload_bytes": 6314, "side_information_bytes": 0, "fetched_bytes": 3439, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26030.547000002116, "ttft_ms": 14.109209063462913, "decode_ms": 95.10904201306403, "receiver_prefill_ms": 335.3455000324175}, "total_wall_ms": 26478.44795801211, "ttft_ms": 14.109209063462913, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4646500", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a4df9595ffaefe76de13f727b9ea78da52720bb08e23db2b848905d48e56b62d", "output_hash": "22c7e5c0df41a8fcf90ee1e6ad966a7b73b0da3ca2ded8274a61b1221ad03f00"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "29f8889dc0000af533ed407f", "attempt_id": "48357391146e", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:01:46.367626Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 2571, "message_tokens": 2479, "message_utf8_bytes": 6342, "cache_tensor_bytes": null, "serialized_payload_bytes": 6342, "side_information_bytes": 0, "fetched_bytes": 3463, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26030.547000002116, "ttft_ms": 13.956999988295138, "decode_ms": 97.7743329713121, "receiver_prefill_ms": 373.1633750721812}, "total_wall_ms": 26519.14200000465, "ttft_ms": 13.956999988295138, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1642260", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "1acffb5aed48fbea8f6057a3f795355392b05d15cc7ad6721cd766ea241bb5b1", "output_hash": "08505337332e43191f2e55bf7664c5072c21fe3054a4238b7a4ab81901c3be71"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "8699ba90164666dcbae60895", "attempt_id": "bf3c9c3ac818", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:01:46.874498Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 2545, "message_tokens": 2444, "message_utf8_bytes": 6285, "cache_tensor_bytes": null, "serialized_payload_bytes": 6285, "side_information_bytes": 0, "fetched_bytes": 3396, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26030.547000002116, "ttft_ms": 13.78512498922646, "decode_ms": 95.24458297528327, "receiver_prefill_ms": 317.5242920406163}, "total_wall_ms": 26461.46066696383, "ttft_ms": 13.78512498922646, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Obysellith-16", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "9641446b74f416e78a8f64bdc714e93c9f99b02b43d752b4b715ff9ed737532f", "output_hash": "66b39064cfe9c103d8574722e2dac0e9a19ced28c716ed2067d84dd8796e7762"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "834f3f8fd8bf7894ce361fe1", "attempt_id": "06c74c1bb216", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:01:47.324044Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 2559, "message_tokens": 2466, "message_utf8_bytes": 6335, "cache_tensor_bytes": null, "serialized_payload_bytes": 6335, "side_information_bytes": 0, "fetched_bytes": 3453, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26030.547000002116, "ttft_ms": 13.782334048300982, "decode_ms": 94.86208390444517, "receiver_prefill_ms": 298.0765000684187}, "total_wall_ms": 26440.559250069782, "ttft_ms": 13.782334048300982, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8069575", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "1e32dc1fd0ee6028872df2d7dd4a31fba388584cd55b62398beeef64edbb1460", "output_hash": "679847bbaae968a3b194b27e6ed62f5fa92be2560b33961650af2f5f21a31e6b"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "50fecbef0abb53d66187735b", "attempt_id": "6e3f16ba1ac1", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:01:47.751973Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 2535, "message_tokens": 2442, "message_utf8_bytes": 6316, "cache_tensor_bytes": null, "serialized_payload_bytes": 6316, "side_information_bytes": 0, "fetched_bytes": 3415, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26030.547000002116, "ttft_ms": 13.902291073463857, "decode_ms": 13.917166041210294, "receiver_prefill_ms": 323.0179159436375}, "total_wall_ms": 26384.69058298506, "ttft_ms": 13.902291073463857, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "5d2bbe281f74a4b6696cb068cb5eaa5e7ade71d3510b0fbab40f0211383df937", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "2828b958497750918e46342c", "attempt_id": "0f5effc0166a", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:01:48.124668Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 2561, "message_tokens": 2466, "message_utf8_bytes": 6335, "cache_tensor_bytes": null, "serialized_payload_bytes": 6335, "side_information_bytes": 0, "fetched_bytes": 3453, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26030.547000002116, "ttft_ms": 14.274165965616703, "decode_ms": 13.697999995201826, "receiver_prefill_ms": 341.50233294349164}, "total_wall_ms": 26403.34729105234, "ttft_ms": 14.274165965616703, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ee7b00506075ba6c6ce4d4e6322c35156472fe10304c803a529c0ace771ddf7b", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "e4f0753f3e4372bb697d1b78", "attempt_id": "8db781cd798a", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:01:49.506924Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 239910912}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 42, "message_tokens": null, "message_utf8_bytes": 30968, "cache_tensor_bytes": 239910912, "serialized_payload_bytes": 239910912, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 990.3491669101641, "ttft_ms": 53.22416697163135, "decode_ms": 110.1440410129726, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1161.667374894023, "ttft_ms": 53.22416697163135, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "9660295", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "91988b1215586dd6899ef1d2a4168b0a54c0112d95ddff92bc07d79f5ac5ee86"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "049bafb0d5fbfd5e940d39bd", "attempt_id": "c9e679eead03", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:01:49.697465Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 239910912}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 46, "message_tokens": null, "message_utf8_bytes": 30968, "cache_tensor_bytes": 239910912, "serialized_payload_bytes": 239910912, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 990.3491669101641, "ttft_ms": 50.353792030364275, "decode_ms": 167.54941700492054, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1212.8624588949606, "ttft_ms": 50.353792030364275, "input_usage": {}, "output_usage": {"output_tokens": 11}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "YF-41T-2802", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "2c942bb83ffbc6ef73a6f94c0666c0c3cc5be2520af627409601e33b0571552b"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "5bd0b21815a5308c6f7fc3d1", "attempt_id": "97f9d065ee4f", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:01:49.939937Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 239910912}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 54, "message_tokens": null, "message_utf8_bytes": 30968, "cache_tensor_bytes": 239910912, "serialized_payload_bytes": 239910912, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 990.3491669101641, "ttft_ms": 89.34691699687392, "decode_ms": 106.0759169049561, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1190.1406669057906, "ttft_ms": 89.34691699687392, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7067400", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "7f470ea83ee0eb2427026d799503d40b32454ebe1ef64f12c8d1a4a53830a524"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "4ecc9078f59fc376ee3cf12e", "attempt_id": "18d209adebf8", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:01:50.158452Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 239910912}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 46, "message_tokens": null, "message_utf8_bytes": 30968, "cache_tensor_bytes": 239910912, "serialized_payload_bytes": 239910912, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 990.3491669101641, "ttft_ms": 49.66616595629603, "decode_ms": 104.94470794219524, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1149.1076248930767, "ttft_ms": 49.66616595629603, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4646500", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "22c7e5c0df41a8fcf90ee1e6ad966a7b73b0da3ca2ded8274a61b1221ad03f00"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "3bc333190983c936176b0b7b", "attempt_id": "ab754856c61e", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:01:50.335335Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 239910912}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 55, "message_tokens": null, "message_utf8_bytes": 30968, "cache_tensor_bytes": 239910912, "serialized_payload_bytes": 239910912, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 990.3491669101641, "ttft_ms": 95.99741699639708, "decode_ms": 105.63516698312014, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1196.5870829299092, "ttft_ms": 95.99741699639708, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Obysellith-16", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "66b39064cfe9c103d8574722e2dac0e9a19ced28c716ed2067d84dd8796e7762"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "591ac70fbce6407326e1c33d", "attempt_id": "5d05c6e993cd", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:01:50.559841Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 239910912}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 30968, "cache_tensor_bytes": 239910912, "serialized_payload_bytes": 239910912, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 990.3491669101641, "ttft_ms": 49.517542007379234, "decode_ms": 105.1483330084011, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1149.6448329417035, "ttft_ms": 49.517542007379234, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "3399575", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "bdc2840939c12a92b36cc897d1b4d1d6fba23d2bd3354af4559cf217aba1973d"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "d460dd06cfd8ae78597fbf18", "attempt_id": "0999929b979f", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:01:50.736992Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 239910912}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 30968, "cache_tensor_bytes": 239910912, "serialized_payload_bytes": 239910912, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 990.3491669101641, "ttft_ms": 49.538624938577414, "decode_ms": 15.123084071092308, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1059.551499900408, "ttft_ms": 49.538624938577414, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "5ab30bf20337dcad6e52130e", "attempt_id": "92dfbcbb61ee", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:01:50.824109Z", "task_family": "synthetic", "document_id": "syn_03004_8000", "document_hash": "9f5712249838ec2b73ad1d3dc0f47c148daf5210f80f5a314286aa23e84ab35d", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 239910912}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5876, "query_tokens": 49, "message_tokens": null, "message_utf8_bytes": 30968, "cache_tensor_bytes": 239910912, "serialized_payload_bytes": 239910912, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 990.3491669101641, "ttft_ms": 49.96299999766052, "decode_ms": 15.12541703414172, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1060.0944998441264, "ttft_ms": 49.96299999766052, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Novostmel-1", "bogus_amount": "8,667,465", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03004_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "691a78776f91f63baefa733b", "attempt_id": "0da33993eeb6", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:01:51.842181Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 2280, "message_tokens": 2194, "message_utf8_bytes": 8048, "cache_tensor_bytes": null, "serialized_payload_bytes": 8048, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.037416000850498676, "ttft_ms": 13.410417013801634, "decode_ms": 122.83679202664644, "receiver_prefill_ms": 258.7785419309512}, "total_wall_ms": 399.00841598864645, "ttft_ms": 13.410417013801634, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5,631,925", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "89f84f1ec8cafcab4e60522da3a7098ddb95248348066ba21f288ca3a96f5b1b", "output_hash": "878945125d6da3defd51b9fd8f53c4954b8aa80d94b40b2fca77c8b7f9a7d2a0"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c213748618f043aa4216b723", "attempt_id": "4842f1451cf4", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:01:52.255988Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 2283, "message_tokens": 2194, "message_utf8_bytes": 8048, "cache_tensor_bytes": null, "serialized_payload_bytes": 8048, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.037416000850498676, "ttft_ms": 13.808040996082127, "decode_ms": 137.1439170325175, "receiver_prefill_ms": 247.97512497752905}, "total_wall_ms": 402.45804109144956, "ttft_ms": 13.808040996082127, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "EP-52G-1330", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "89f84f1ec8cafcab4e60522da3a7098ddb95248348066ba21f288ca3a96f5b1b", "output_hash": "104db207937fd79e23ce16f9b3f10b1a9ed34f6dd0441a99be6ea1b2d0a09d99"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "e1824d8f25170fad5ba05296", "attempt_id": "a14ea9cbed2f", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:01:52.673536Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 2288, "message_tokens": 2194, "message_utf8_bytes": 8048, "cache_tensor_bytes": null, "serialized_payload_bytes": 8048, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.037416000850498676, "ttft_ms": 13.494124985300004, "decode_ms": 106.44708306062967, "receiver_prefill_ms": 247.14433297049254}, "total_wall_ms": 370.6472909543663, "ttft_ms": 13.494124985300004, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "10107545", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "89f84f1ec8cafcab4e60522da3a7098ddb95248348066ba21f288ca3a96f5b1b", "output_hash": "6faba9de4e29949bbdd8934f69fc2110f07076db6801b211aeca9351292537fe"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "09a6e6b01c2806d388cb4f96", "attempt_id": "238d9aa7ad30", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:01:53.059016Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 2281, "message_tokens": 2194, "message_utf8_bytes": 8048, "cache_tensor_bytes": null, "serialized_payload_bytes": 8048, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.037416000850498676, "ttft_ms": 13.495624996721745, "decode_ms": 94.27612496074289, "receiver_prefill_ms": 247.44920898228884}, "total_wall_ms": 358.71533304452896, "ttft_ms": 13.495624996721745, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8474840", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "89f84f1ec8cafcab4e60522da3a7098ddb95248348066ba21f288ca3a96f5b1b", "output_hash": "e7f39dffe6d5859c740d1a7ee67b1f6f8997d2c087ce7d2da9ad0eef582f32c1"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b26feb14cdb949903ae37efc", "attempt_id": "db6efc29b05d", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:01:53.432123Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 2290, "message_tokens": 2194, "message_utf8_bytes": 8048, "cache_tensor_bytes": null, "serialized_payload_bytes": 8048, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.037416000850498676, "ttft_ms": 13.495541992597282, "decode_ms": 81.43174997530878, "receiver_prefill_ms": 247.52795801032335}, "total_wall_ms": 346.0913320304826, "ttft_ms": 13.495541992597282, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Praymi-11", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "89f84f1ec8cafcab4e60522da3a7098ddb95248348066ba21f288ca3a96f5b1b", "output_hash": "849597882f995af2b27ff5870d3e46a2cdcc4c3f2814b3873f634447a51e5f26"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b3f490f693edd3f52a5d6b60", "attempt_id": "ef45d905aa89", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:01:53.793143Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 2283, "message_tokens": 2194, "message_utf8_bytes": 8048, "cache_tensor_bytes": null, "serialized_payload_bytes": 8048, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.037416000850498676, "ttft_ms": 13.21895804721862, "decode_ms": 94.29720800835639, "receiver_prefill_ms": 247.65983398538083}, "total_wall_ms": 358.6177070392296, "ttft_ms": 13.21895804721862, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4186578", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "89f84f1ec8cafcab4e60522da3a7098ddb95248348066ba21f288ca3a96f5b1b", "output_hash": "3ad53922762faa5b31b197e89b5e37ab1cbcb19ad482a131723c26a8a88defd6"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "4d9f0385d5d1301af037f60e", "attempt_id": "5808a6133cd1", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:01:54.165884Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 2284, "message_tokens": 2194, "message_utf8_bytes": 8048, "cache_tensor_bytes": null, "serialized_payload_bytes": 8048, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.037416000850498676, "ttft_ms": 13.229333912022412, "decode_ms": 13.459375011734664, "receiver_prefill_ms": 247.16949998401105}, "total_wall_ms": 277.34470705036074, "ttft_ms": 13.229333912022412, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "89f84f1ec8cafcab4e60522da3a7098ddb95248348066ba21f288ca3a96f5b1b", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "9eb9ccb5a668007d4cf6e96e", "attempt_id": "a4e6d38addcb", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:01:54.457574Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 2285, "message_tokens": 2194, "message_utf8_bytes": 8048, "cache_tensor_bytes": null, "serialized_payload_bytes": 8048, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.037416000850498676, "ttft_ms": 13.421500101685524, "decode_ms": 13.432124978862703, "receiver_prefill_ms": 248.18541598506272}, "total_wall_ms": 278.4807499265298, "ttft_ms": 13.421500101685524, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "89f84f1ec8cafcab4e60522da3a7098ddb95248348066ba21f288ca3a96f5b1b", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "728a8a98f1c435d3464c1d24", "attempt_id": "e7755c7854ce", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:02:08.740257Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 1125, "message_tokens": 1037, "message_utf8_bytes": 2469, "cache_tensor_bytes": null, "serialized_payload_bytes": 2469, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13990.123457973823, "ttft_ms": 13.296458055265248, "decode_ms": 120.71420799475163, "receiver_prefill_ms": 121.95562501437962}, "total_wall_ms": 14247.965415939689, "ttft_ms": 13.296458055265248, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5,631,925", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "4aa2cd4ec4096e5ba7828d8aebcb7c7283b082e79be8224a59b0d4638cd5cc7b", "output_hash": "878945125d6da3defd51b9fd8f53c4954b8aa80d94b40b2fca77c8b7f9a7d2a0"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "78958071cdef46cbf195f46c", "attempt_id": "a1b704cafbc2", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:02:09.013177Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 1128, "message_tokens": 1037, "message_utf8_bytes": 2469, "cache_tensor_bytes": null, "serialized_payload_bytes": 2469, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13990.123457973823, "ttft_ms": 13.250165968202055, "decode_ms": 133.12595803290606, "receiver_prefill_ms": 121.67716701515019}, "total_wall_ms": 14260.023499955423, "ttft_ms": 13.250165968202055, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "EP-52G-1330", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "4aa2cd4ec4096e5ba7828d8aebcb7c7283b082e79be8224a59b0d4638cd5cc7b", "output_hash": "104db207937fd79e23ce16f9b3f10b1a9ed34f6dd0441a99be6ea1b2d0a09d99"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "ec3481987d4d9ec5e7d5f056", "attempt_id": "5469ac144505", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:02:09.297803Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 1133, "message_tokens": 1037, "message_utf8_bytes": 2469, "cache_tensor_bytes": null, "serialized_payload_bytes": 2469, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13990.123457973823, "ttft_ms": 13.781249988824129, "decode_ms": 106.868541915901, "receiver_prefill_ms": 121.54887500219047}, "total_wall_ms": 14234.169748960994, "ttft_ms": 13.781249988824129, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "10107545", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "4aa2cd4ec4096e5ba7828d8aebcb7c7283b082e79be8224a59b0d4638cd5cc7b", "output_hash": "6faba9de4e29949bbdd8934f69fc2110f07076db6801b211aeca9351292537fe"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "f6c454dc1da6d61836652433", "attempt_id": "ccec7b835870", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:02:09.556739Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 1126, "message_tokens": 1037, "message_utf8_bytes": 2469, "cache_tensor_bytes": null, "serialized_payload_bytes": 2469, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13990.123457973823, "ttft_ms": 13.318250072188675, "decode_ms": 94.08370801247656, "receiver_prefill_ms": 121.39016599394381}, "total_wall_ms": 14220.668375026435, "ttft_ms": 13.318250072188675, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8474840", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "4aa2cd4ec4096e5ba7828d8aebcb7c7283b082e79be8224a59b0d4638cd5cc7b", "output_hash": "e7f39dffe6d5859c740d1a7ee67b1f6f8997d2c087ce7d2da9ad0eef582f32c1"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "8578e45e8a8f710b5fb8d8a0", "attempt_id": "5b726526d3f6", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:02:09.801426Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 1135, "message_tokens": 1037, "message_utf8_bytes": 2469, "cache_tensor_bytes": null, "serialized_payload_bytes": 2469, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13990.123457973823, "ttft_ms": 13.695959001779556, "decode_ms": 80.00766695477068, "receiver_prefill_ms": 121.08512490522116}, "total_wall_ms": 14206.808998947963, "ttft_ms": 13.695959001779556, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Praymi-11", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "4aa2cd4ec4096e5ba7828d8aebcb7c7283b082e79be8224a59b0d4638cd5cc7b", "output_hash": "849597882f995af2b27ff5870d3e46a2cdcc4c3f2814b3873f634447a51e5f26"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "fafaf715edbabba02678a78f", "attempt_id": "bd69a264597a", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:02:10.032034Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 1128, "message_tokens": 1037, "message_utf8_bytes": 2469, "cache_tensor_bytes": null, "serialized_payload_bytes": 2469, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13990.123457973823, "ttft_ms": 13.52012506686151, "decode_ms": 133.39129195082933, "receiver_prefill_ms": 122.89587501436472}, "total_wall_ms": 14261.750207981095, "ttft_ms": 13.52012506686151, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "14,590,283", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "4aa2cd4ec4096e5ba7828d8aebcb7c7283b082e79be8224a59b0d4638cd5cc7b", "output_hash": "85d342e3fc9c80d7739fa2e444eb82a759d2f2b9b3c893edba55ab66d904f0c6"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "7ca09a8acb1935fb9e71c801", "attempt_id": "0afa0f0481a7", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:02:10.318776Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 1129, "message_tokens": 1037, "message_utf8_bytes": 2469, "cache_tensor_bytes": null, "serialized_payload_bytes": 2469, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13990.123457973823, "ttft_ms": 13.53895803913474, "decode_ms": 13.436374953016639, "receiver_prefill_ms": 127.32558394782245}, "total_wall_ms": 14146.157040959224, "ttft_ms": 13.53895803913474, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "4aa2cd4ec4096e5ba7828d8aebcb7c7283b082e79be8224a59b0d4638cd5cc7b", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "fec6a795ada5daf7e18518d2", "attempt_id": "4f263310102d", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:02:10.489449Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 1130, "message_tokens": 1037, "message_utf8_bytes": 2469, "cache_tensor_bytes": null, "serialized_payload_bytes": 2469, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13990.123457973823, "ttft_ms": 13.792083016596735, "decode_ms": 13.36966606322676, "receiver_prefill_ms": 133.41945898719132}, "total_wall_ms": 14152.50999899581, "ttft_ms": 13.792083016596735, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "4aa2cd4ec4096e5ba7828d8aebcb7c7283b082e79be8224a59b0d4638cd5cc7b", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "63654935390c08fadb21e752", "attempt_id": "be466a9838f7", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:02:24.599882Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 16}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 717, "message_tokens": 626, "message_utf8_bytes": 1618, "cache_tensor_bytes": null, "serialized_payload_bytes": 1618, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13933.650332968682, "ttft_ms": 13.162959017790854, "decode_ms": 13.436709064990282, "receiver_prefill_ms": 78.11666699126363}, "total_wall_ms": 14039.998832973652, "ttft_ms": 13.162959017790854, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "9125e93f36682b46ce6accfa4af6316044460bbafd3bd8e6075ada8b8e1b59eb", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "85f1b5291a2c9ffe6c61e28d", "attempt_id": "d349655c1b6d", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:02:24.720594Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 16}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 720, "message_tokens": 626, "message_utf8_bytes": 1618, "cache_tensor_bytes": null, "serialized_payload_bytes": 1618, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13933.650332968682, "ttft_ms": 13.113624998368323, "decode_ms": 13.168749981559813, "receiver_prefill_ms": 78.17762508057058}, "total_wall_ms": 14039.401207934134, "ttft_ms": 13.113624998368323, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "9125e93f36682b46ce6accfa4af6316044460bbafd3bd8e6075ada8b8e1b59eb", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "9067d12a37b528b631cd81db", "attempt_id": "90afba7798ec", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:02:24.841184Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 16}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 725, "message_tokens": 626, "message_utf8_bytes": 1618, "cache_tensor_bytes": null, "serialized_payload_bytes": 1618, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13933.650332968682, "ttft_ms": 13.223999994806945, "decode_ms": 105.86404101923108, "receiver_prefill_ms": 78.08937504887581}, "total_wall_ms": 14132.195833022706, "ttft_ms": 13.223999994806945, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "10107545", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "9125e93f36682b46ce6accfa4af6316044460bbafd3bd8e6075ada8b8e1b59eb", "output_hash": "6faba9de4e29949bbdd8934f69fc2110f07076db6801b211aeca9351292537fe"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "aea512e8a2e4593b91b20069", "attempt_id": "b3a3149448c8", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:02:25.054455Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 16}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 718, "message_tokens": 626, "message_utf8_bytes": 1618, "cache_tensor_bytes": null, "serialized_payload_bytes": 1618, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13933.650332968682, "ttft_ms": 13.063165941275656, "decode_ms": 132.78420898132026, "receiver_prefill_ms": 77.81695807352662}, "total_wall_ms": 14158.69075001683, "ttft_ms": 13.063165941275656, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "10,107,545", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "9125e93f36682b46ce6accfa4af6316044460bbafd3bd8e6075ada8b8e1b59eb", "output_hash": "3d6a277763361ef55b8dd3f9791012605d24cd167ac5bdcdd0064ee16de161d4"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "eaab24134b9f848a55fb93db", "attempt_id": "84100e6fbddd", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:02:25.293563Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 16}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 727, "message_tokens": 626, "message_utf8_bytes": 1618, "cache_tensor_bytes": null, "serialized_payload_bytes": 1618, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13933.650332968682, "ttft_ms": 13.171916012652218, "decode_ms": 78.86529096867889, "receiver_prefill_ms": 78.34474998526275}, "total_wall_ms": 14105.362207978033, "ttft_ms": 13.171916012652218, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Praymi-11", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "9125e93f36682b46ce6accfa4af6316044460bbafd3bd8e6075ada8b8e1b59eb", "output_hash": "849597882f995af2b27ff5870d3e46a2cdcc4c3f2814b3873f634447a51e5f26"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "5352cf7935e8c20f85674bdf", "attempt_id": "95897bb6f4a3", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:02:25.479527Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 16}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 720, "message_tokens": 626, "message_utf8_bytes": 1618, "cache_tensor_bytes": null, "serialized_payload_bytes": 1618, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13933.650332968682, "ttft_ms": 13.409958919510245, "decode_ms": 105.24587496183813, "receiver_prefill_ms": 77.60258298367262}, "total_wall_ms": 14131.226832978427, "ttft_ms": 13.409958919510245, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "11615.75", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "9125e93f36682b46ce6accfa4af6316044460bbafd3bd8e6075ada8b8e1b59eb", "output_hash": "459c141f1960a681ed5a178c354c50e5bb5347e1a8a542d2925cf0b5cfc0a3b9"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "dc67a8976814929f4d4bb6c6", "attempt_id": "bc67014f28e6", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:02:25.691444Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 16}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 721, "message_tokens": 626, "message_utf8_bytes": 1618, "cache_tensor_bytes": null, "serialized_payload_bytes": 1618, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13933.650332968682, "ttft_ms": 13.34512501489371, "decode_ms": 13.523375033400953, "receiver_prefill_ms": 78.28566699754447}, "total_wall_ms": 14040.110374917276, "ttft_ms": 13.34512501489371, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "9125e93f36682b46ce6accfa4af6316044460bbafd3bd8e6075ada8b8e1b59eb", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "8d1b55e9e25eb77638d76d6c", "attempt_id": "b0b4fcdf6dc2", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:02:25.812221Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 16}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 722, "message_tokens": 626, "message_utf8_bytes": 1618, "cache_tensor_bytes": null, "serialized_payload_bytes": 1618, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13933.650332968682, "ttft_ms": 13.421833049505949, "decode_ms": 13.126584002748132, "receiver_prefill_ms": 77.93933292850852}, "total_wall_ms": 14039.429249009117, "ttft_ms": 13.421833049505949, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "9125e93f36682b46ce6accfa4af6316044460bbafd3bd8e6075ada8b8e1b59eb", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "62705fbb574daf823ffb8f9f", "attempt_id": "e9004fb80872", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:02:39.788742Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 748, "message_tokens": 658, "message_utf8_bytes": 1732, "cache_tensor_bytes": null, "serialized_payload_bytes": 1732, "side_information_bytes": 0, "fetched_bytes": 851, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13857.400374952704, "ttft_ms": 13.19274993147701, "decode_ms": 13.337999931536615, "receiver_prefill_ms": 81.83633303269744}, "total_wall_ms": 13967.638917034492, "ttft_ms": 13.19274993147701, "input_usage": {"lookups": 3}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "3cb5f7504837901e1413e0cb93826b5dd86753c10ee34f32282c9a19028a8692", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "56ac4622ddd6780f39316377", "attempt_id": "8543a1673160", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:02:39.917043Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 709, "message_tokens": 616, "message_utf8_bytes": 1647, "cache_tensor_bytes": null, "serialized_payload_bytes": 1647, "side_information_bytes": 0, "fetched_bytes": 770, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13857.400374952704, "ttft_ms": 13.589374953880906, "decode_ms": 13.208416057750583, "receiver_prefill_ms": 77.83216703683138}, "total_wall_ms": 13963.504749932326, "ttft_ms": 13.589374953880906, "input_usage": {"lookups": 3}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "963707b4f665835b85c22f2b26dd9550e40098b9d36b7afd09e1fe9e6780fa67", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "ad645dc1295fea8359687eec", "attempt_id": "bcf7b945e571", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:02:40.037856Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 756, "message_tokens": 658, "message_utf8_bytes": 1732, "cache_tensor_bytes": null, "serialized_payload_bytes": 1732, "side_information_bytes": 0, "fetched_bytes": 851, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13857.400374952704, "ttft_ms": 13.365291990339756, "decode_ms": 13.190499972552061, "receiver_prefill_ms": 81.50262502022088}, "total_wall_ms": 13966.872499906458, "ttft_ms": 13.365291990339756, "input_usage": {"lookups": 3}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ebc8b9c08869fe5a2c016b30ff200db4bb630778225243ca4c0657a086104d73", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "fd81fd0cb29928a81a08f8ce", "attempt_id": "d9bce4b41b1e", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:02:40.161790Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 749, "message_tokens": 658, "message_utf8_bytes": 1732, "cache_tensor_bytes": null, "serialized_payload_bytes": 1732, "side_information_bytes": 0, "fetched_bytes": 851, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13857.400374952704, "ttft_ms": 13.087208033539355, "decode_ms": 13.423583004623652, "receiver_prefill_ms": 81.0276250122115}, "total_wall_ms": 13966.324749984778, "ttft_ms": 13.087208033539355, "input_usage": {"lookups": 3}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ebc8b9c08869fe5a2c016b30ff200db4bb630778225243ca4c0657a086104d73", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "dddaa038f40392521a8e913a", "attempt_id": "13b582c4e644", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:02:40.285157Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 752, "message_tokens": 652, "message_utf8_bytes": 1733, "cache_tensor_bytes": null, "serialized_payload_bytes": 1733, "side_information_bytes": 0, "fetched_bytes": 848, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13857.400374952704, "ttft_ms": 13.16158298868686, "decode_ms": 13.238082989118993, "receiver_prefill_ms": 82.30079093482345}, "total_wall_ms": 13967.523249913938, "ttft_ms": 13.16158298868686, "input_usage": {"lookups": 3}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "4ed7d0d8287b8605b8f5ac264ac7f24a5ed1c33a297e3439fd2ed96a67bb8967", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "d61c82230ccb36e6a281d271", "attempt_id": "3ed76730d2a1", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:02:40.409833Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 751, "message_tokens": 658, "message_utf8_bytes": 1732, "cache_tensor_bytes": null, "serialized_payload_bytes": 1732, "side_information_bytes": 0, "fetched_bytes": 851, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13857.400374952704, "ttft_ms": 13.391457963734865, "decode_ms": 92.81083301175386, "receiver_prefill_ms": 80.86962508969009}, "total_wall_ms": 14045.912457862869, "ttft_ms": 13.391457963734865, "input_usage": {"lookups": 3}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8908553", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ebc8b9c08869fe5a2c016b30ff200db4bb630778225243ca4c0657a086104d73", "output_hash": "5ad9e777dc3be36808c74badfc8aae217b41a486754d07adf1abda6f6bce9aca"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "ea2bdecf60e18f88600da653", "attempt_id": "ff1bfa86beb8", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:02:40.612662Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 738, "message_tokens": 644, "message_utf8_bytes": 1720, "cache_tensor_bytes": null, "serialized_payload_bytes": 1720, "side_information_bytes": 0, "fetched_bytes": 833, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13857.400374952704, "ttft_ms": 13.545791967771947, "decode_ms": 13.282959000207484, "receiver_prefill_ms": 80.86670795455575}, "total_wall_ms": 13966.485874960199, "ttft_ms": 13.545791967771947, "input_usage": {"lookups": 3}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "554eb27cbe12dc3f92632ae89582ab604bdc4fc431b41cb715894bf18d11d4e6", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a0c17cb1835e5f181a6c5117", "attempt_id": "8c4f0463f9e8", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:02:40.735286Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 753, "message_tokens": 658, "message_utf8_bytes": 1732, "cache_tensor_bytes": null, "serialized_payload_bytes": 1732, "side_information_bytes": 0, "fetched_bytes": 851, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13857.400374952704, "ttft_ms": 13.64237500820309, "decode_ms": 13.18291702773422, "receiver_prefill_ms": 81.07804099563509}, "total_wall_ms": 13966.805916978046, "ttft_ms": 13.64237500820309, "input_usage": {"lookups": 3}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "3cb5f7504837901e1413e0cb93826b5dd86753c10ee34f32282c9a19028a8692", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "bd356ba243c02b394ae6b95a", "attempt_id": "6f51d22ad65f", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:02:41.140122Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 83755008}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 44, "message_tokens": null, "message_utf8_bytes": 8022, "cache_tensor_bytes": 83755008, "serialized_payload_bytes": 83755008, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 276.8395410384983, "ttft_ms": 30.079875024966896, "decode_ms": 124.95208298787475, "receiver_prefill_ms": 0.0}, "total_wall_ms": 437.69595806952566, "ttft_ms": 30.079875024966896, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5,631,925", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "878945125d6da3defd51b9fd8f53c4954b8aa80d94b40b2fca77c8b7f9a7d2a0"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "3aaf382013dcd0b010e83e23", "attempt_id": "e5aaf823a55c", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:02:41.315821Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 83755008}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 8022, "cache_tensor_bytes": 83755008, "serialized_payload_bytes": 83755008, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 276.8395410384983, "ttft_ms": 30.604333034716547, "decode_ms": 137.9853329854086, "receiver_prefill_ms": 0.0}, "total_wall_ms": 449.10533307120204, "ttft_ms": 30.604333034716547, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "EP-52G-1330", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "104db207937fd79e23ce16f9b3f10b1a9ed34f6dd0441a99be6ea1b2d0a09d99"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "1d24b57865b517a76a2b2d74", "attempt_id": "d741855526dd", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:02:41.505267Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 83755008}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 52, "message_tokens": null, "message_utf8_bytes": 8022, "cache_tensor_bytes": 83755008, "serialized_payload_bytes": 83755008, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 276.8395410384983, "ttft_ms": 27.15899993199855, "decode_ms": 110.37308303639293, "receiver_prefill_ms": 0.0}, "total_wall_ms": 418.27495803590864, "ttft_ms": 27.15899993199855, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "10107545", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "6faba9de4e29949bbdd8934f69fc2110f07076db6801b211aeca9351292537fe"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "5fd885f1a703633e975ad631", "attempt_id": "9db87688f7c9", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:02:41.661044Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 83755008}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 45, "message_tokens": null, "message_utf8_bytes": 8022, "cache_tensor_bytes": 83755008, "serialized_payload_bytes": 83755008, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 276.8395410384983, "ttft_ms": 28.31266599241644, "decode_ms": 13.89433303847909, "receiver_prefill_ms": 0.0}, "total_wall_ms": 322.79308303259313, "ttft_ms": 28.31266599241644, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "7e4fd4c5362778e217994d2f", "attempt_id": "d99a6dcbcf6f", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:02:41.721405Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 83755008}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 54, "message_tokens": null, "message_utf8_bytes": 8022, "cache_tensor_bytes": 83755008, "serialized_payload_bytes": 83755008, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 276.8395410384983, "ttft_ms": 27.636750019155443, "decode_ms": 82.34162500593811, "receiver_prefill_ms": 0.0}, "total_wall_ms": 390.58024901896715, "ttft_ms": 27.636750019155443, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Draosttek-2", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "2b3e2f36e669a543bfead6d94871991d5157079209ba24d18bc5c19f41b9d99e"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "8d2d17cce4d62df8905bcc0c", "attempt_id": "92b3ca257fdf", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:02:41.849516Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 83755008}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 8022, "cache_tensor_bytes": 83755008, "serialized_payload_bytes": 83755008, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 276.8395410384983, "ttft_ms": 27.874708059243858, "decode_ms": 137.6476250588894, "receiver_prefill_ms": 0.0}, "total_wall_ms": 446.1224999977276, "ttft_ms": 27.874708059243858, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "23,345,973", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "904c858f0e325df6a5c26b026fe210e6b6d32e147f565522c15eea2dbb2f472d"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "50232e63f0d9a02baf1d5372", "attempt_id": "0ab9415d1692", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:02:42.032979Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 83755008}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 48, "message_tokens": null, "message_utf8_bytes": 8022, "cache_tensor_bytes": 83755008, "serialized_payload_bytes": 83755008, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 276.8395410384983, "ttft_ms": 27.05187490209937, "decode_ms": 13.60120892059058, "receiver_prefill_ms": 0.0}, "total_wall_ms": 321.15779106970876, "ttft_ms": 27.05187490209937, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "5dc9df1a2fe5245f6e51cf44", "attempt_id": "f45de2e3df80", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:02:42.092707Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 83755008}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 49, "message_tokens": null, "message_utf8_bytes": 8022, "cache_tensor_bytes": 83755008, "serialized_payload_bytes": 83755008, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 276.8395410384983, "ttft_ms": 28.111333958804607, "decode_ms": 13.703917036764324, "receiver_prefill_ms": 0.0}, "total_wall_ms": 322.177249006927, "ttft_ms": 28.111333958804607, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "aafbbdefe478459fee89013b", "attempt_id": "6a9833e887ae", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:02:42.155102Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 2394, "message_tokens": 2308, "message_utf8_bytes": 8552, "cache_tensor_bytes": null, "serialized_payload_bytes": 8552, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.028792070224881172, "ttft_ms": 13.72708403505385, "decode_ms": 120.72795792482793, "receiver_prefill_ms": 264.94316605385393}, "total_wall_ms": 402.998001081869, "ttft_ms": 13.72708403505385, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5,631,925", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "3ea06bbd2f93d7d5a06d1a1bd8d3f4f1d170b499d29f34923b9d54c1a1cc4bee", "output_hash": "878945125d6da3defd51b9fd8f53c4954b8aa80d94b40b2fca77c8b7f9a7d2a0"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a545c9b4f663ff9150fe42f2", "attempt_id": "2f9ecb89016c", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:02:42.572049Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 2397, "message_tokens": 2308, "message_utf8_bytes": 8552, "cache_tensor_bytes": null, "serialized_payload_bytes": 8552, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.028792070224881172, "ttft_ms": 13.792500016279519, "decode_ms": 134.61904204450548, "receiver_prefill_ms": 276.31912496872246}, "total_wall_ms": 428.53537504561245, "ttft_ms": 13.792500016279519, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "EP-52G-1330", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "3ea06bbd2f93d7d5a06d1a1bd8d3f4f1d170b499d29f34923b9d54c1a1cc4bee", "output_hash": "104db207937fd79e23ce16f9b3f10b1a9ed34f6dd0441a99be6ea1b2d0a09d99"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "39c4cc33ae3da11acc94abda", "attempt_id": "029f5973dd13", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:02:43.015016Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 2402, "message_tokens": 2308, "message_utf8_bytes": 8552, "cache_tensor_bytes": null, "serialized_payload_bytes": 8552, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.028792070224881172, "ttft_ms": 13.741582981310785, "decode_ms": 108.31483395304531, "receiver_prefill_ms": 281.28991706762463}, "total_wall_ms": 406.98825009167194, "ttft_ms": 13.741582981310785, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "10107545", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "3ea06bbd2f93d7d5a06d1a1bd8d3f4f1d170b499d29f34923b9d54c1a1cc4bee", "output_hash": "6faba9de4e29949bbdd8934f69fc2110f07076db6801b211aeca9351292537fe"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "0f42e15c7a78a5b271274884", "attempt_id": "cef21bdfcf41", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:02:43.436215Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 2395, "message_tokens": 2308, "message_utf8_bytes": 8552, "cache_tensor_bytes": null, "serialized_payload_bytes": 8552, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.028792070224881172, "ttft_ms": 13.400209019891918, "decode_ms": 94.78520904667675, "receiver_prefill_ms": 274.76341708097607}, "total_wall_ms": 386.60254213027656, "ttft_ms": 13.400209019891918, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8474840", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "3ea06bbd2f93d7d5a06d1a1bd8d3f4f1d170b499d29f34923b9d54c1a1cc4bee", "output_hash": "e7f39dffe6d5859c740d1a7ee67b1f6f8997d2c087ce7d2da9ad0eef582f32c1"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "72fd60233addb08c99f0bd50", "attempt_id": "7b4c473ad9cd", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:02:43.837257Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 2404, "message_tokens": 2308, "message_utf8_bytes": 8552, "cache_tensor_bytes": null, "serialized_payload_bytes": 8552, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.028792070224881172, "ttft_ms": 13.766749994829297, "decode_ms": 81.39908302109689, "receiver_prefill_ms": 283.3237500162795}, "total_wall_ms": 382.3504999745637, "ttft_ms": 13.766749994829297, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Praymi-11", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "3ea06bbd2f93d7d5a06d1a1bd8d3f4f1d170b499d29f34923b9d54c1a1cc4bee", "output_hash": "849597882f995af2b27ff5870d3e46a2cdcc4c3f2814b3873f634447a51e5f26"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "236cb911586b8f9186cd1dc3", "attempt_id": "ef449fe8c827", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:02:44.237527Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 2397, "message_tokens": 2308, "message_utf8_bytes": 8552, "cache_tensor_bytes": null, "serialized_payload_bytes": 8552, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.028792070224881172, "ttft_ms": 13.640500023029745, "decode_ms": 95.25658295024186, "receiver_prefill_ms": 284.1729159699753}, "total_wall_ms": 396.6674170223996, "ttft_ms": 13.640500023029745, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4044458", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "3ea06bbd2f93d7d5a06d1a1bd8d3f4f1d170b499d29f34923b9d54c1a1cc4bee", "output_hash": "3521b22667520edd85e15fa99fcb80828a392a164728c1f9374a90f642205bf2"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c0bbad03072ad2e014accd3c", "attempt_id": "098479b60131", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:02:44.649126Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 2398, "message_tokens": 2308, "message_utf8_bytes": 8552, "cache_tensor_bytes": null, "serialized_payload_bytes": 8552, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.028792070224881172, "ttft_ms": 13.543999986723065, "decode_ms": 13.46708310302347, "receiver_prefill_ms": 285.52820801269263}, "total_wall_ms": 316.1591669777408, "ttft_ms": 13.543999986723065, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "3ea06bbd2f93d7d5a06d1a1bd8d3f4f1d170b499d29f34923b9d54c1a1cc4bee", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "ae845309633b34af6b68dfcc", "attempt_id": "dcc7c28ba315", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:02:44.979544Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 2399, "message_tokens": 2308, "message_utf8_bytes": 8552, "cache_tensor_bytes": null, "serialized_payload_bytes": 8552, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.028792070224881172, "ttft_ms": 13.685125042684376, "decode_ms": 13.72854202054441, "receiver_prefill_ms": 295.8814590238035}, "total_wall_ms": 326.88412512652576, "ttft_ms": 13.685125042684376, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "3ea06bbd2f93d7d5a06d1a1bd8d3f4f1d170b499d29f34923b9d54c1a1cc4bee", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "6dd7601264c5f73ce49e5c7d", "attempt_id": "47fd62f1f4fb", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:03:07.619124Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1647}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 1758, "message_tokens": 1671, "message_utf8_bytes": 5336, "cache_tensor_bytes": null, "serialized_payload_bytes": 5336, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 22298.30754105933, "ttft_ms": 13.366375002078712, "decode_ms": 119.57004200667143, "receiver_prefill_ms": 183.09762503486127}, "total_wall_ms": 22619.031832087785, "ttft_ms": 13.366375002078712, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5,631,925", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "84830d0d3d9d25622a344549ecbb40fdbc836763a30a7412e779884bd4e0be0e", "output_hash": "878945125d6da3defd51b9fd8f53c4954b8aa80d94b40b2fca77c8b7f9a7d2a0"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "9f085e85fd1c2e26741d8d0f", "attempt_id": "1474d8d43e70", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:03:07.954828Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1647}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 1761, "message_tokens": 1671, "message_utf8_bytes": 5336, "cache_tensor_bytes": null, "serialized_payload_bytes": 5336, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 22298.30754105933, "ttft_ms": 13.330708025023341, "decode_ms": 133.58004193287343, "receiver_prefill_ms": 181.17029091808945}, "total_wall_ms": 22629.0888331132, "ttft_ms": 13.330708025023341, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "EP-52G-1330", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "84830d0d3d9d25622a344549ecbb40fdbc836763a30a7412e779884bd4e0be0e", "output_hash": "104db207937fd79e23ce16f9b3f10b1a9ed34f6dd0441a99be6ea1b2d0a09d99"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b728395146f2d2469c8292d8", "attempt_id": "c3517e92d690", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:03:08.302844Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1647}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 1766, "message_tokens": 1671, "message_utf8_bytes": 5336, "cache_tensor_bytes": null, "serialized_payload_bytes": 5336, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 22298.30754105933, "ttft_ms": 13.736083055846393, "decode_ms": 106.52029095217586, "receiver_prefill_ms": 185.87787495926023}, "total_wall_ms": 22607.177666039206, "ttft_ms": 13.736083055846393, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "10107545", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "84830d0d3d9d25622a344549ecbb40fdbc836763a30a7412e779884bd4e0be0e", "output_hash": "6faba9de4e29949bbdd8934f69fc2110f07076db6801b211aeca9351292537fe"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "6077621e204f4a61453bbcf6", "attempt_id": "47a4b4a3dbfe", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:03:08.626089Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1647}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 1759, "message_tokens": 1671, "message_utf8_bytes": 5336, "cache_tensor_bytes": null, "serialized_payload_bytes": 5336, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 22298.30754105933, "ttft_ms": 13.778166961856186, "decode_ms": 94.11554108373821, "receiver_prefill_ms": 189.7631250321865}, "total_wall_ms": 22598.609082051553, "ttft_ms": 13.778166961856186, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "9112735", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "84830d0d3d9d25622a344549ecbb40fdbc836763a30a7412e779884bd4e0be0e", "output_hash": "15a3f804636f0e62d4a3d9be7677f001a0ad6b1fb8ae7e8cdc63c44ec7bac31f"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "292b7437ed11e64790893475", "attempt_id": "e4ad48c3e449", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:03:08.940793Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1647}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 1768, "message_tokens": 1671, "message_utf8_bytes": 5336, "cache_tensor_bytes": null, "serialized_payload_bytes": 5336, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 22298.30754105933, "ttft_ms": 13.79791705403477, "decode_ms": 81.09641703777015, "receiver_prefill_ms": 207.7699579531327}, "total_wall_ms": 22603.703499073163, "ttft_ms": 13.79791705403477, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Praymi-11", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "84830d0d3d9d25622a344549ecbb40fdbc836763a30a7412e779884bd4e0be0e", "output_hash": "849597882f995af2b27ff5870d3e46a2cdcc4c3f2814b3873f634447a51e5f26"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "e2a6d8149040b14c6c6dff2a", "attempt_id": "64c26f2ee8c9", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:03:09.261284Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1647}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 1761, "message_tokens": 1671, "message_utf8_bytes": 5336, "cache_tensor_bytes": null, "serialized_payload_bytes": 5336, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 22298.30754105933, "ttft_ms": 13.53408396244049, "decode_ms": 94.34979199431837, "receiver_prefill_ms": 204.44891694933176}, "total_wall_ms": 22613.29458304681, "ttft_ms": 13.53408396244049, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4084878", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "84830d0d3d9d25622a344549ecbb40fdbc836763a30a7412e779884bd4e0be0e", "output_hash": "9107a710ae0db0290d06d2f3c52367d3ed99d14568e945685a8f0bc59d656987"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "fbc1ff23e7a3b62551453852", "attempt_id": "a277634a53e4", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:03:09.590541Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1647}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 1762, "message_tokens": 1671, "message_utf8_bytes": 5336, "cache_tensor_bytes": null, "serialized_payload_bytes": 5336, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 22298.30754105933, "ttft_ms": 13.621458085253835, "decode_ms": 13.73312494251877, "receiver_prefill_ms": 208.3015829557553}, "total_wall_ms": 22536.770791048184, "ttft_ms": 13.621458085253835, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "84830d0d3d9d25622a344549ecbb40fdbc836763a30a7412e779884bd4e0be0e", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "5919cf12fd22fc8f28395b96", "attempt_id": "84bf0fe23878", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:03:09.843671Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1647}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 1763, "message_tokens": 1671, "message_utf8_bytes": 5336, "cache_tensor_bytes": null, "serialized_payload_bytes": 5336, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 22298.30754105933, "ttft_ms": 13.889666995964944, "decode_ms": 13.541458989493549, "receiver_prefill_ms": 217.3295000102371}, "total_wall_ms": 22545.876375050284, "ttft_ms": 13.889666995964944, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "84830d0d3d9d25622a344549ecbb40fdbc836763a30a7412e779884bd4e0be0e", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "0971aeb229a7895cc7f9bf49", "attempt_id": "eaacebb3cf82", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:03:37.745787Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 31}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 1284, "message_tokens": 1193, "message_utf8_bytes": 3177, "cache_tensor_bytes": null, "serialized_payload_bytes": 3177, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27638.572917087004, "ttft_ms": 13.522624969482422, "decode_ms": 13.475915999151766, "receiver_prefill_ms": 139.90745798218995}, "total_wall_ms": 27807.392625138164, "ttft_ms": 13.522624969482422, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "b357b617ac77797d52c34a7e23e68ebc82a1f827ac0a6d975ee4b20ef28ef59d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "1aa0307e29b233c6460153cf", "attempt_id": "2d3b9bdc790f", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:03:37.934525Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 31}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 1287, "message_tokens": 1193, "message_utf8_bytes": 3177, "cache_tensor_bytes": null, "serialized_payload_bytes": 3177, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27638.572917087004, "ttft_ms": 13.704792014323175, "decode_ms": 13.299374957568944, "receiver_prefill_ms": 222.69233304541558}, "total_wall_ms": 27890.22937498521, "ttft_ms": 13.704792014323175, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "b357b617ac77797d52c34a7e23e68ebc82a1f827ac0a6d975ee4b20ef28ef59d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "1a0d5cddf6711de9a7fbd9d3", "attempt_id": "01a0ac6a249a", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:03:38.201015Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 31}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 1292, "message_tokens": 1193, "message_utf8_bytes": 3177, "cache_tensor_bytes": null, "serialized_payload_bytes": 3177, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27638.572917087004, "ttft_ms": 14.226665953174233, "decode_ms": 116.24566605314612, "receiver_prefill_ms": 627.5459160096943}, "total_wall_ms": 28398.73495802749, "ttft_ms": 14.226665953174233, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "10107545", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "b357b617ac77797d52c34a7e23e68ebc82a1f827ac0a6d975ee4b20ef28ef59d", "output_hash": "6faba9de4e29949bbdd8934f69fc2110f07076db6801b211aeca9351292537fe"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "4585bf43bc1decbdc0bfb357", "attempt_id": "1d840e5c79c3", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:03:38.976173Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 31}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 1285, "message_tokens": 1193, "message_utf8_bytes": 3177, "cache_tensor_bytes": null, "serialized_payload_bytes": 3177, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27638.572917087004, "ttft_ms": 13.505791896022856, "decode_ms": 107.1995000820607, "receiver_prefill_ms": 140.41154191363603}, "total_wall_ms": 27901.768750161864, "ttft_ms": 13.505791896022856, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "10107545", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "b357b617ac77797d52c34a7e23e68ebc82a1f827ac0a6d975ee4b20ef28ef59d", "output_hash": "6faba9de4e29949bbdd8934f69fc2110f07076db6801b211aeca9351292537fe"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "3e0cf1e3fa295335146dd64d", "attempt_id": "56794fed032b", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:03:39.293616Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 31}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 1294, "message_tokens": 1193, "message_utf8_bytes": 3177, "cache_tensor_bytes": null, "serialized_payload_bytes": 3177, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27638.572917087004, "ttft_ms": 13.960958109237254, "decode_ms": 81.38300001155585, "receiver_prefill_ms": 802.1567919058725}, "total_wall_ms": 28537.96458407305, "ttft_ms": 13.960958109237254, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Draosttek-2", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "b357b617ac77797d52c34a7e23e68ebc82a1f827ac0a6d975ee4b20ef28ef59d", "output_hash": "2b3e2f36e669a543bfead6d94871991d5157079209ba24d18bc5c19f41b9d99e"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "4c74a2743781b31079c396b0", "attempt_id": "8dbd7e459a8a", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:03:40.207180Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 31}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 1287, "message_tokens": 1193, "message_utf8_bytes": 3177, "cache_tensor_bytes": null, "serialized_payload_bytes": 3177, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27638.572917087004, "ttft_ms": 13.62508290912956, "decode_ms": 134.42033401224762, "receiver_prefill_ms": 164.92324997670949}, "total_wall_ms": 27953.549959114753, "ttft_ms": 13.62508290912956, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "26,865,348", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "b357b617ac77797d52c34a7e23e68ebc82a1f827ac0a6d975ee4b20ef28ef59d", "output_hash": "96317a115d507530ebe6269ced3eb771a8b27b28c76134cc67e33c9863f8ba7d"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "ddeae121bc20054b93d1525d", "attempt_id": "51e31289c1a8", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:03:40.537587Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 31}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 1288, "message_tokens": 1193, "message_utf8_bytes": 3177, "cache_tensor_bytes": null, "serialized_payload_bytes": 3177, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27638.572917087004, "ttft_ms": 13.64837505389005, "decode_ms": 13.492707978002727, "receiver_prefill_ms": 140.62175003346056}, "total_wall_ms": 27808.289542095736, "ttft_ms": 13.64837505389005, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "b357b617ac77797d52c34a7e23e68ebc82a1f827ac0a6d975ee4b20ef28ef59d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "43c02fc210a184eedfe65faf", "attempt_id": "feea08b8cbd9", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:03:40.721568Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 31}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 1289, "message_tokens": 1193, "message_utf8_bytes": 3177, "cache_tensor_bytes": null, "serialized_payload_bytes": 3177, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27638.572917087004, "ttft_ms": 13.922083075158298, "decode_ms": 13.73241690453142, "receiver_prefill_ms": 203.60195799730718}, "total_wall_ms": 27871.83420918882, "ttft_ms": 13.922083075158298, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "b357b617ac77797d52c34a7e23e68ebc82a1f827ac0a6d975ee4b20ef28ef59d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "904777176f3887a804e1c759", "attempt_id": "5f69673c4c16", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:04:08.733975Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 2731, "message_tokens": 2641, "message_utf8_bytes": 6759, "cache_tensor_bytes": null, "serialized_payload_bytes": 6759, "side_information_bytes": 0, "fetched_bytes": 3468, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27764.86645801924, "ttft_ms": 14.155374956317246, "decode_ms": 13.727499986998737, "receiver_prefill_ms": 420.71787500754}, "total_wall_ms": 28218.38437509723, "ttft_ms": 14.155374956317246, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "e7f95cfb92d03913e154c6f8b970c2842c8f3c1f50b4f8d6b1a753d2f375b3c7", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "00c5afb9ad10d17e65bd7500", "attempt_id": "ae89efb170fa", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:04:09.206103Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 2627, "message_tokens": 2534, "message_utf8_bytes": 6437, "cache_tensor_bytes": null, "serialized_payload_bytes": 6437, "side_information_bytes": 0, "fetched_bytes": 3173, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27764.86645801924, "ttft_ms": 13.672000030055642, "decode_ms": 133.5417500231415, "receiver_prefill_ms": 287.02699998393655}, "total_wall_ms": 28206.273583113216, "ttft_ms": 13.672000030055642, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "PX-79Z-3050", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "c0f447aeca1e7b2831043ef19da1fd403b9ebb25227296d256e8435b01143375", "output_hash": "4aab49dea099f1d98c6fddf46de1f70be98609e58a7b5a0fa6b1ed0e7320a24b"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "40a73d09f74a68d9ecfa35c0", "attempt_id": "fc1a0eba41bb", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:04:09.662928Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 2732, "message_tokens": 2634, "message_utf8_bytes": 6728, "cache_tensor_bytes": null, "serialized_payload_bytes": 6728, "side_information_bytes": 0, "fetched_bytes": 3436, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27764.86645801924, "ttft_ms": 14.351458987221122, "decode_ms": 111.75887496210635, "receiver_prefill_ms": 359.60612492635846}, "total_wall_ms": 28254.2252499843, "ttft_ms": 14.351458987221122, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "10107545", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "e5f53a9859ba355c81f0a92d8fe256a8639c041047cc5aaf632e6f4bc61860af", "output_hash": "6faba9de4e29949bbdd8934f69fc2110f07076db6801b211aeca9351292537fe"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "10913cc74638bb0f5c3e86c7", "attempt_id": "eb7241b0fba2", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:04:10.166752Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 2725, "message_tokens": 2634, "message_utf8_bytes": 6728, "cache_tensor_bytes": null, "serialized_payload_bytes": 6728, "side_information_bytes": 0, "fetched_bytes": 3436, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27764.86645801924, "ttft_ms": 13.65225005429238, "decode_ms": 109.12183299660683, "receiver_prefill_ms": 339.5959170302376}, "total_wall_ms": 28231.51254095137, "ttft_ms": 13.65225005429238, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "10107545", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "f6be7a289558537cc09de2158fe1bfef0c0264d598058af680a88381a54f5964", "output_hash": "6faba9de4e29949bbdd8934f69fc2110f07076db6801b211aeca9351292537fe"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "4b55ae9653d8b55ed85f0cba", "attempt_id": "d2f0979a4ae8", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:04:10.647930Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 2639, "message_tokens": 2539, "message_utf8_bytes": 6444, "cache_tensor_bytes": null, "serialized_payload_bytes": 6444, "side_information_bytes": 0, "fetched_bytes": 3197, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27764.86645801924, "ttft_ms": 14.337416971102357, "decode_ms": 83.10683304443955, "receiver_prefill_ms": 366.34758301079273}, "total_wall_ms": 28232.383791008033, "ttft_ms": 14.337416971102357, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Draosttek-2", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "b3ea663f0126bc51c205727cbb9b9b13ce5985e3ad7214dea6d1e1ac876f6daa", "output_hash": "2b3e2f36e669a543bfead6d94871991d5157079209ba24d18bc5c19f41b9d99e"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "e6ff6a4faa2ffb22ec408258", "attempt_id": "21b371a6b3bd", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:04:11.129904Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 2736, "message_tokens": 2643, "message_utf8_bytes": 6785, "cache_tensor_bytes": null, "serialized_payload_bytes": 6785, "side_information_bytes": 0, "fetched_bytes": 3485, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27764.86645801924, "ttft_ms": 14.08387499395758, "decode_ms": 137.05712498631328, "receiver_prefill_ms": 324.4161659386009}, "total_wall_ms": 28244.25266601611, "ttft_ms": 14.08387499395758, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "29,830,088", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "637c3ea63db647865cdb0a31e6ddddbcf388425a79e190164d8e73e392266b0f", "output_hash": "8ac188172016babf916811c66fa8398ccd9c8d69ff82c95682352ff02a26f036"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "fbecb7860df5b3c5b9d67337", "attempt_id": "4a65b884b498", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:04:11.625579Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 2698, "message_tokens": 2604, "message_utf8_bytes": 6655, "cache_tensor_bytes": null, "serialized_payload_bytes": 6655, "side_information_bytes": 0, "fetched_bytes": 3359, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27764.86645801924, "ttft_ms": 13.806332950480282, "decode_ms": 14.204875100404024, "receiver_prefill_ms": 321.6164589393884}, "total_wall_ms": 28118.034292012453, "ttft_ms": 13.806332950480282, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "51440997c97fb44b02bf073925e4f6cac043ac2efba28dbc271b70ecca0f7f03", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "19255ac954a66c770379c077", "attempt_id": "0e1f626175d3", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:04:11.993123Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 2715, "message_tokens": 2620, "message_utf8_bytes": 6678, "cache_tensor_bytes": null, "serialized_payload_bytes": 6678, "side_information_bytes": 0, "fetched_bytes": 3394, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27764.86645801924, "ttft_ms": 14.115291996859014, "decode_ms": 14.519624994136393, "receiver_prefill_ms": 412.3917910037562}, "total_wall_ms": 28209.533542045392, "ttft_ms": 14.115291996859014, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "e8173af4bad0cd6a41f8b8bcccc96dd3f7e5ead4fa090d7cb72d4a9c40dabd86", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "324ae164645290bc694c35c2", "attempt_id": "571bb18347a3", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:04:13.254551Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 90832896}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 44, "message_tokens": null, "message_utf8_bytes": 8498, "cache_tensor_bytes": 90832896, "serialized_payload_bytes": 90832896, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 802.3782500531524, "ttft_ms": 356.31108295638114, "decode_ms": 129.4249159982428, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1292.2322921222076, "ttft_ms": 356.31108295638114, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5,631,925", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "878945125d6da3defd51b9fd8f53c4954b8aa80d94b40b2fca77c8b7f9a7d2a0"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "9196b8691ebc9d055d7ca22c", "attempt_id": "e2f3609fe183", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:04:13.760252Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 90832896}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 8498, "cache_tensor_bytes": 90832896, "serialized_payload_bytes": 90832896, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 802.3782500531524, "ttft_ms": 378.71937500312924, "decode_ms": 142.13166595436633, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1325.8593330392614, "ttft_ms": 378.71937500312924, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "UM-25F-5156", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "fd70c1f0818207943d7ead36d04df74d8bebd6d8c465337df8086eb0f1c07e81"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "5d6dceb3f606e15d6ff2a7f2", "attempt_id": "305240c0671c", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:04:14.298458Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 90832896}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 52, "message_tokens": null, "message_utf8_bytes": 8498, "cache_tensor_bytes": 90832896, "serialized_payload_bytes": 90832896, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 802.3782500531524, "ttft_ms": 29.906542040407658, "decode_ms": 102.15637495275587, "receiver_prefill_ms": 0.0}, "total_wall_ms": 937.2005419572815, "ttft_ms": 29.906542040407658, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "9766130", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "6a300110eb4bcf2805c784a6f7ecd53011c195bd1530514e09cd9eb70bedd00a"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "021abad7cbac467db75b67d6", "attempt_id": "a37c9e0c180f", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:04:14.448839Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 90832896}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 45, "message_tokens": null, "message_utf8_bytes": 8498, "cache_tensor_bytes": 90832896, "serialized_payload_bytes": 90832896, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 802.3782500531524, "ttft_ms": 30.222625005990267, "decode_ms": 14.676040969789028, "receiver_prefill_ms": 0.0}, "total_wall_ms": 849.9424160690978, "ttft_ms": 30.222625005990267, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "d7097417cc88a8a204e2aaaf", "attempt_id": "b85e1914ae50", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:04:14.513667Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 90832896}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 54, "message_tokens": null, "message_utf8_bytes": 8498, "cache_tensor_bytes": 90832896, "serialized_payload_bytes": 90832896, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 802.3782500531524, "ttft_ms": 30.01016692724079, "decode_ms": 86.88645798247308, "receiver_prefill_ms": 0.0}, "total_wall_ms": 922.0755830174312, "ttft_ms": 30.01016692724079, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Draosttek-2", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "2b3e2f36e669a543bfead6d94871991d5157079209ba24d18bc5c19f41b9d99e"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a4acab296096fbd216eca7b3", "attempt_id": "ffac4989c76c", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:04:14.650539Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 90832896}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 8498, "cache_tensor_bytes": 90832896, "serialized_payload_bytes": 90832896, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 802.3782500531524, "ttft_ms": 29.438458033837378, "decode_ms": 114.28629199508578, "receiver_prefill_ms": 0.0}, "total_wall_ms": 948.7331659765914, "ttft_ms": 29.438458033837378, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "15131275", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "23d59e84823ca437d2b93ee98227aa1661c664d90cbc00a346bee0f99a70d281"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "46830f6ba59dcb57e586b285", "attempt_id": "f68a4da43975", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:04:14.812064Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 90832896}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 48, "message_tokens": null, "message_utf8_bytes": 8498, "cache_tensor_bytes": 90832896, "serialized_payload_bytes": 90832896, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 802.3782500531524, "ttft_ms": 28.787666000425816, "decode_ms": 14.242334058508277, "receiver_prefill_ms": 0.0}, "total_wall_ms": 848.1041251216084, "ttft_ms": 28.787666000425816, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c0e8a06dd56446cbfc313702", "attempt_id": "c0f37410058f", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:04:14.872098Z", "task_family": "synthetic", "document_id": "syn_03006_2000", "document_hash": "89be5e77e0ca114963e9ffdf8dc38fe5610f13a78ac29e7a1956cd9124ce5217", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 90832896}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1890, "query_tokens": 49, "message_tokens": null, "message_utf8_bytes": 8498, "cache_tensor_bytes": 90832896, "serialized_payload_bytes": 90832896, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 802.3782500531524, "ttft_ms": 29.678332968614995, "decode_ms": 14.345625066198409, "receiver_prefill_ms": 0.0}, "total_wall_ms": 849.0745420567691, "ttft_ms": 29.678332968614995, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Praymi-11", "bogus_amount": "8,474,840", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03006_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "bb293e10b6303ec8c52e02c8", "attempt_id": "7456d2c94117", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:04:16.024059Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 6222, "message_tokens": 6136, "message_utf8_bytes": 30068, "cache_tensor_bytes": null, "serialized_payload_bytes": 30068, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.03191700670868158, "ttft_ms": 14.263249933719635, "decode_ms": 96.96116705890745, "receiver_prefill_ms": 744.6186660090461}, "total_wall_ms": 866.7382089188322, "ttft_ms": 14.263249933719635, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4150875", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S1"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "86155f66d84c85a93a5e81e34b6cb071057e40c178ac0fec1d2e6bb5de9e5c97", "output_hash": "9f908cf585946283f90239132608e9f36a14f3587bfad3942073463dff2947a8"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c1969544bb4acb748d5bc305", "attempt_id": "2f670eae31f6", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:04:16.909077Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 6224, "message_tokens": 6136, "message_utf8_bytes": 30068, "cache_tensor_bytes": null, "serialized_payload_bytes": 30068, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.03191700670868158, "ttft_ms": 13.725167023949325, "decode_ms": 136.5813750308007, "receiver_prefill_ms": 730.5301249725744}, "total_wall_ms": 891.4154170779511, "ttft_ms": 13.725167023949325, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "BY-78E-8447", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S1"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "86155f66d84c85a93a5e81e34b6cb071057e40c178ac0fec1d2e6bb5de9e5c97", "output_hash": "6acdf1f1807eeb525a30ee2d553ebf18ba3bb86ebe5cef8081992457efd3bb8d"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "819d76464dc46e85366e72a5", "attempt_id": "4539790f5542", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:04:17.817338Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 6230, "message_tokens": 6136, "message_utf8_bytes": 30068, "cache_tensor_bytes": null, "serialized_payload_bytes": 30068, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.03191700670868158, "ttft_ms": 13.86529195588082, "decode_ms": 95.46970797237009, "receiver_prefill_ms": 729.5461249304935}, "total_wall_ms": 848.5668339999393, "ttft_ms": 13.86529195588082, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5403805", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S1"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "86155f66d84c85a93a5e81e34b6cb071057e40c178ac0fec1d2e6bb5de9e5c97", "output_hash": "ebd8c202e450e10a6d90509f598803c77ba6ed9c928c98df1eaddfa99eb333a4"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "4ae1331afb298df9c66ff771", "attempt_id": "c13180711ea7", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:04:18.683317Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 6223, "message_tokens": 6136, "message_utf8_bytes": 30068, "cache_tensor_bytes": null, "serialized_payload_bytes": 30068, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.03191700670868158, "ttft_ms": 13.946958933956921, "decode_ms": 96.16258391179144, "receiver_prefill_ms": 744.4961669389158}, "total_wall_ms": 864.5440420368686, "ttft_ms": 13.946958933956921, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4150875", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S1"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "86155f66d84c85a93a5e81e34b6cb071057e40c178ac0fec1d2e6bb5de9e5c97", "output_hash": "9f908cf585946283f90239132608e9f36a14f3587bfad3942073463dff2947a8"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "918565ec93b1c15dfacd3041", "attempt_id": "80f6db802985", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:04:19.564791Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 6233, "message_tokens": 6136, "message_utf8_bytes": 30068, "cache_tensor_bytes": null, "serialized_payload_bytes": 30068, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.03191700670868158, "ttft_ms": 13.669417006894946, "decode_ms": 96.16895799990743, "receiver_prefill_ms": 765.1250830385834}, "total_wall_ms": 884.9852089770138, "ttft_ms": 13.669417006894946, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Keryminov-10", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S1"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "86155f66d84c85a93a5e81e34b6cb071057e40c178ac0fec1d2e6bb5de9e5c97", "output_hash": "47d4fb882fdd2db10f1e88942b3f981e24a10a4a4501ff56cdaf321c9eba1c18"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c65362d5a1f948fc89da0c01", "attempt_id": "0970e62c2702", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:04:20.466974Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 6225, "message_tokens": 6136, "message_utf8_bytes": 30068, "cache_tensor_bytes": null, "serialized_payload_bytes": 30068, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.03191700670868158, "ttft_ms": 13.992083026096225, "decode_ms": 96.29316697828472, "receiver_prefill_ms": 788.1045829271898}, "total_wall_ms": 908.6086259922013, "ttft_ms": 13.992083026096225, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "2328859", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S1"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "86155f66d84c85a93a5e81e34b6cb071057e40c178ac0fec1d2e6bb5de9e5c97", "output_hash": "be43d88b9cdee93dc31161cbc51446019dddf20a3dd715826d3e6b81e4518672"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "03325e013b1ac81e37eb1ef5", "attempt_id": "26ae89bf4724", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:04:21.393242Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 6228, "message_tokens": 6136, "message_utf8_bytes": 30068, "cache_tensor_bytes": null, "serialized_payload_bytes": 30068, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.03191700670868158, "ttft_ms": 13.82566592656076, "decode_ms": 13.865791028365493, "receiver_prefill_ms": 808.2444580504671}, "total_wall_ms": 845.7475421018898, "ttft_ms": 13.82566592656076, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S1"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "86155f66d84c85a93a5e81e34b6cb071057e40c178ac0fec1d2e6bb5de9e5c97", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "4944d5cfeb3f8c11babed988", "attempt_id": "4e313b7200c3", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:04:22.256868Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 6227, "message_tokens": 6136, "message_utf8_bytes": 30068, "cache_tensor_bytes": null, "serialized_payload_bytes": 30068, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.03191700670868158, "ttft_ms": 13.86379194445908, "decode_ms": 13.88991589192301, "receiver_prefill_ms": 850.2620840445161}, "total_wall_ms": 887.7541250549257, "ttft_ms": 13.86379194445908, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S1"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "86155f66d84c85a93a5e81e34b6cb071057e40c178ac0fec1d2e6bb5de9e5c97", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "9bd96de0826a72b7573ef949", "attempt_id": "0c1b902c1520", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:04:37.737421Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 1125, "message_tokens": 1037, "message_utf8_bytes": 2339, "cache_tensor_bytes": null, "serialized_payload_bytes": 2339, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14575.288249994628, "ttft_ms": 13.82829190697521, "decode_ms": 93.99237495381385, "receiver_prefill_ms": 121.81920802686363}, "total_wall_ms": 14806.852291920222, "ttft_ms": 13.82829190697521, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4644474", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S1"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "15594ec3a9ee3fc5073b27d8dd49921db28f65cc66ecef5d13fded2dac1f15aa", "output_hash": "89ba2585f859d43ddc46bdb0cb238c294f76729603c976374065611b0682c797"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "02aa34177f8919de07861699", "attempt_id": "2fb5aa4bdd20", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:04:37.987109Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 1127, "message_tokens": 1037, "message_utf8_bytes": 2339, "cache_tensor_bytes": null, "serialized_payload_bytes": 2339, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14575.288249994628, "ttft_ms": 13.510416960343719, "decode_ms": 133.23662499897182, "receiver_prefill_ms": 121.45849992521107}, "total_wall_ms": 14845.306666917168, "ttft_ms": 13.510416960343719, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "BY-78E-8447", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S1"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "15594ec3a9ee3fc5073b27d8dd49921db28f65cc66ecef5d13fded2dac1f15aa", "output_hash": "6acdf1f1807eeb525a30ee2d553ebf18ba3bb86ebe5cef8081992457efd3bb8d"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "10ea0ae8f75839cfe4010585", "attempt_id": "99db4e2e2ce0", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:04:38.274530Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 1133, "message_tokens": 1037, "message_utf8_bytes": 2339, "cache_tensor_bytes": null, "serialized_payload_bytes": 2339, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14575.288249994628, "ttft_ms": 13.504792004823685, "decode_ms": 94.14929198101163, "receiver_prefill_ms": 121.39166705310345}, "total_wall_ms": 14806.119250017218, "ttft_ms": 13.504792004823685, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5403805", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S1"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "15594ec3a9ee3fc5073b27d8dd49921db28f65cc66ecef5d13fded2dac1f15aa", "output_hash": "ebd8c202e450e10a6d90509f598803c77ba6ed9c928c98df1eaddfa99eb333a4"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "6d6fa285d68def80765162b1", "attempt_id": "37b4ed4cef4b", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:04:38.522598Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 1126, "message_tokens": 1037, "message_utf8_bytes": 2339, "cache_tensor_bytes": null, "serialized_payload_bytes": 2339, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14575.288249994628, "ttft_ms": 13.462083064951003, "decode_ms": 93.49145798478276, "receiver_prefill_ms": 121.19320803321898}, "total_wall_ms": 14805.25541701354, "ttft_ms": 13.462083064951003, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5403805", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S1"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "15594ec3a9ee3fc5073b27d8dd49921db28f65cc66ecef5d13fded2dac1f15aa", "output_hash": "ebd8c202e450e10a6d90509f598803c77ba6ed9c928c98df1eaddfa99eb333a4"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "9ea1766fab85f1d1cff56e8e", "attempt_id": "eca128797929", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:04:38.770217Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 1136, "message_tokens": 1037, "message_utf8_bytes": 2339, "cache_tensor_bytes": null, "serialized_payload_bytes": 2339, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14575.288249994628, "ttft_ms": 13.593166950158775, "decode_ms": 94.2648749332875, "receiver_prefill_ms": 121.6419170377776}, "total_wall_ms": 14806.623249896802, "ttft_ms": 13.593166950158775, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Keryminov-10", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S1"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "15594ec3a9ee3fc5073b27d8dd49921db28f65cc66ecef5d13fded2dac1f15aa", "output_hash": "47d4fb882fdd2db10f1e88942b3f981e24a10a4a4501ff56cdaf321c9eba1c18"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "0b8f277e8a6e40678277f5a1", "attempt_id": "dc0fe1430087", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:04:39.018479Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 1128, "message_tokens": 1037, "message_utf8_bytes": 2339, "cache_tensor_bytes": null, "serialized_payload_bytes": 2339, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14575.288249994628, "ttft_ms": 13.355250004678965, "decode_ms": 93.50266703404486, "receiver_prefill_ms": 121.44670798443258}, "total_wall_ms": 14805.446416023187, "ttft_ms": 13.355250004678965, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1648464", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S1"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "15594ec3a9ee3fc5073b27d8dd49921db28f65cc66ecef5d13fded2dac1f15aa", "output_hash": "8a08f184ece3105d60bcf50de406c7bb8946dd104cd4654660d43b6d0b2f70c5"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "87b2466013389daecc7ef36f", "attempt_id": "30212c96db31", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:04:39.265815Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 1131, "message_tokens": 1037, "message_utf8_bytes": 2339, "cache_tensor_bytes": null, "serialized_payload_bytes": 2339, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14575.288249994628, "ttft_ms": 13.437582994811237, "decode_ms": 13.395666028372943, "receiver_prefill_ms": 121.27079104539007}, "total_wall_ms": 14725.126083008945, "ttft_ms": 13.437582994811237, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S1"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "15594ec3a9ee3fc5073b27d8dd49921db28f65cc66ecef5d13fded2dac1f15aa", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "e800bc2f861ec981854e0f31", "attempt_id": "b1cb2a076688", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:04:39.432661Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 1130, "message_tokens": 1037, "message_utf8_bytes": 2339, "cache_tensor_bytes": null, "serialized_payload_bytes": 2339, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14575.288249994628, "ttft_ms": 13.754458050243556, "decode_ms": 13.466250034980476, "receiver_prefill_ms": 123.02116595674306}, "total_wall_ms": 14727.29916695971, "ttft_ms": 13.754458050243556, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S1"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "15594ec3a9ee3fc5073b27d8dd49921db28f65cc66ecef5d13fded2dac1f15aa", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "662046fa441a77b66dcf3b11", "attempt_id": "35fad69d2196", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:04:51.427032Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 14}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 643, "message_tokens": 552, "message_utf8_bytes": 1550, "cache_tensor_bytes": null, "serialized_payload_bytes": 1550, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11825.126208015718, "ttft_ms": 13.138040900230408, "decode_ms": 13.289332971908152, "receiver_prefill_ms": 75.15570800751448}, "total_wall_ms": 11930.823873961344, "ttft_ms": 13.138040900230408, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S1"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ad347a935b2eb2abc849ac257f19791975135739e0d0a3bde9249decfbf04e3d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "8511452aeadd6b07ce529a91", "attempt_id": "08ea6ba51993", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:04:51.550556Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 14}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 645, "message_tokens": 552, "message_utf8_bytes": 1550, "cache_tensor_bytes": null, "serialized_payload_bytes": 1550, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11825.126208015718, "ttft_ms": 13.493624981492758, "decode_ms": 132.3725830297917, "receiver_prefill_ms": 74.8604170512408}, "total_wall_ms": 12047.306665917858, "ttft_ms": 13.493624981492758, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "BY-78E-8447", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S1"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ad347a935b2eb2abc849ac257f19791975135739e0d0a3bde9249decfbf04e3d", "output_hash": "6acdf1f1807eeb525a30ee2d553ebf18ba3bb86ebe5cef8081992457efd3bb8d"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "dbc251fd317fab2dad6bce74", "attempt_id": "8960ca13bd6d", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:04:51.791192Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 14}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 651, "message_tokens": 552, "message_utf8_bytes": 1550, "cache_tensor_bytes": null, "serialized_payload_bytes": 1550, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11825.126208015718, "ttft_ms": 13.181042042560875, "decode_ms": 13.128624996170402, "receiver_prefill_ms": 74.89933306351304}, "total_wall_ms": 11927.637165994383, "ttft_ms": 13.181042042560875, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S1"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ad347a935b2eb2abc849ac257f19791975135739e0d0a3bde9249decfbf04e3d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "66235141b332025d5df5823e", "attempt_id": "f924ed3364d7", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:04:51.910735Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 14}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 644, "message_tokens": 552, "message_utf8_bytes": 1550, "cache_tensor_bytes": null, "serialized_payload_bytes": 1550, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11825.126208015718, "ttft_ms": 13.3382499916479, "decode_ms": 13.057249947451055, "receiver_prefill_ms": 74.55745793413371}, "total_wall_ms": 11927.327833021991, "ttft_ms": 13.3382499916479, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S1"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ad347a935b2eb2abc849ac257f19791975135739e0d0a3bde9249decfbf04e3d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "74ba0093d015f794b2e2cc46", "attempt_id": "9a06a41974db", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:04:52.029605Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 14}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 654, "message_tokens": 552, "message_utf8_bytes": 1550, "cache_tensor_bytes": null, "serialized_payload_bytes": 1550, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11825.126208015718, "ttft_ms": 13.170124962925911, "decode_ms": 78.82462500128895, "receiver_prefill_ms": 74.82158299535513}, "total_wall_ms": 11993.22487495374, "ttft_ms": 13.170124962925911, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Lithixymi-6", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S1"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ad347a935b2eb2abc849ac257f19791975135739e0d0a3bde9249decfbf04e3d", "output_hash": "5919268fa866f12eb8b0497583904be835032b475afd544dc2eba31f2b4f6679"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "8ebb472e2ad0840a950199d5", "attempt_id": "72b734e25dd0", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:04:52.215097Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 14}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 646, "message_tokens": 552, "message_utf8_bytes": 1550, "cache_tensor_bytes": null, "serialized_payload_bytes": 1550, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11825.126208015718, "ttft_ms": 13.301083003170788, "decode_ms": 13.298999983817339, "receiver_prefill_ms": 74.59387497510761}, "total_wall_ms": 11927.59820795618, "ttft_ms": 13.301083003170788, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S1"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ad347a935b2eb2abc849ac257f19791975135739e0d0a3bde9249decfbf04e3d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "7d2fb67a7c3d3d6009679c3d", "attempt_id": "ee9ce5403dc2", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:04:52.335103Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 14}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 649, "message_tokens": 552, "message_utf8_bytes": 1550, "cache_tensor_bytes": null, "serialized_payload_bytes": 1550, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11825.126208015718, "ttft_ms": 13.415792025625706, "decode_ms": 13.912416994571686, "receiver_prefill_ms": 74.69104102347046}, "total_wall_ms": 11928.59408305958, "ttft_ms": 13.415792025625706, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S1"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ad347a935b2eb2abc849ac257f19791975135739e0d0a3bde9249decfbf04e3d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b9dc425a7e3c8bbf6ed6fc2a", "attempt_id": "2977c0ac8af7", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:04:52.455387Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 14}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 648, "message_tokens": 552, "message_utf8_bytes": 1550, "cache_tensor_bytes": null, "serialized_payload_bytes": 1550, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11825.126208015718, "ttft_ms": 13.133457978256047, "decode_ms": 13.185458024963737, "receiver_prefill_ms": 74.73120803479105}, "total_wall_ms": 11927.622791961767, "ttft_ms": 13.133457978256047, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S1"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ad347a935b2eb2abc849ac257f19791975135739e0d0a3bde9249decfbf04e3d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "7c6706d633d37d63b8da80ef", "attempt_id": "e0bb46bf550d", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:05:04.421440Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 657, "message_tokens": 567, "message_utf8_bytes": 1429, "cache_tensor_bytes": null, "serialized_payload_bytes": 1429, "side_information_bytes": 0, "fetched_bytes": 809, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11846.203958033584, "ttft_ms": 13.30270804464817, "decode_ms": 13.113750028423965, "receiver_prefill_ms": 74.93004098068923}, "total_wall_ms": 11948.889000108466, "ttft_ms": 13.30270804464817, "input_usage": {"lookups": 3}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S1"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "8f98f10ad14c4a28016ef3eebb001913df39a6fa6ad55da93f968d1cd1f1eedf", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "8e33bee44bd8e2628692a9b8", "attempt_id": "55044ef3f5ae", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:05:04.541758Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 659, "message_tokens": 567, "message_utf8_bytes": 1429, "cache_tensor_bytes": null, "serialized_payload_bytes": 1429, "side_information_bytes": 0, "fetched_bytes": 809, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11846.203958033584, "ttft_ms": 13.441249961033463, "decode_ms": 13.372665969654918, "receiver_prefill_ms": 74.87504207529128}, "total_wall_ms": 11949.235250009224, "ttft_ms": 13.441249961033463, "input_usage": {"lookups": 3}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S1"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "8f98f10ad14c4a28016ef3eebb001913df39a6fa6ad55da93f968d1cd1f1eedf", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "ec9805b6e11a212fe2ff3bea", "attempt_id": "f552f93c0eda", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:05:04.662394Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 665, "message_tokens": 567, "message_utf8_bytes": 1429, "cache_tensor_bytes": null, "serialized_payload_bytes": 1429, "side_information_bytes": 0, "fetched_bytes": 809, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11846.203958033584, "ttft_ms": 13.156707980670035, "decode_ms": 13.255666941404343, "receiver_prefill_ms": 75.06958395242691}, "total_wall_ms": 11949.010540964082, "ttft_ms": 13.156707980670035, "input_usage": {"lookups": 3}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S1"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "8f98f10ad14c4a28016ef3eebb001913df39a6fa6ad55da93f968d1cd1f1eedf", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "34b5e813b68adfc94bebc1ef", "attempt_id": "d8db87aa2d20", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:05:04.782651Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 658, "message_tokens": 567, "message_utf8_bytes": 1429, "cache_tensor_bytes": null, "serialized_payload_bytes": 1429, "side_information_bytes": 0, "fetched_bytes": 809, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11846.203958033584, "ttft_ms": 13.369875028729439, "decode_ms": 93.00466696731746, "receiver_prefill_ms": 74.556541047059}, "total_wall_ms": 12028.595999930985, "ttft_ms": 13.369875028729439, "input_usage": {"lookups": 3}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "6456275", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S1"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "8f98f10ad14c4a28016ef3eebb001913df39a6fa6ad55da93f968d1cd1f1eedf", "output_hash": "8459afe0e14fc306ba5922b08df60f937337ca27b185956d38f7d94627fd5349"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "f1f06436397345dc2364e50c", "attempt_id": "0f668a27a1b4", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:05:04.982341Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 668, "message_tokens": 567, "message_utf8_bytes": 1429, "cache_tensor_bytes": null, "serialized_payload_bytes": 1429, "side_information_bytes": 0, "fetched_bytes": 809, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11846.203958033584, "ttft_ms": 13.443415984511375, "decode_ms": 13.116167043335736, "receiver_prefill_ms": 75.06587507668883}, "total_wall_ms": 11949.203917058185, "ttft_ms": 13.443415984511375, "input_usage": {"lookups": 3}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S1"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "d1430b692dacc23bc9d262a6bbae067665e7a566f24fcec3de1cf3ed82fe4e40", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "cc4f052105c1f3527e56ddac", "attempt_id": "f5d9eae59b18", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:05:05.102598Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 660, "message_tokens": 567, "message_utf8_bytes": 1429, "cache_tensor_bytes": null, "serialized_payload_bytes": 1429, "side_information_bytes": 0, "fetched_bytes": 809, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11846.203958033584, "ttft_ms": 13.283042004331946, "decode_ms": 132.71470798645169, "receiver_prefill_ms": 74.7592500410974}, "total_wall_ms": 12068.317916011438, "ttft_ms": 13.283042004331946, "input_usage": {"lookups": 3}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "15,045,503", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S1"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "8f98f10ad14c4a28016ef3eebb001913df39a6fa6ad55da93f968d1cd1f1eedf", "output_hash": "a16f61f15041df32151ae39622681709040d0e88063fd10266f93a28fec7a87e"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "7e0f3b9d3c1de338687eb8d4", "attempt_id": "8eddb0e97eae", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:05:05.341912Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 663, "message_tokens": 567, "message_utf8_bytes": 1429, "cache_tensor_bytes": null, "serialized_payload_bytes": 1429, "side_information_bytes": 0, "fetched_bytes": 809, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11846.203958033584, "ttft_ms": 13.120459043420851, "decode_ms": 13.297208002768457, "receiver_prefill_ms": 74.53216705471277}, "total_wall_ms": 11948.429041076452, "ttft_ms": 13.120459043420851, "input_usage": {"lookups": 3}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S1"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "8f98f10ad14c4a28016ef3eebb001913df39a6fa6ad55da93f968d1cd1f1eedf", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c306dcb7fac5c7b3c9afe699", "attempt_id": "e12feb9e75af", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:05:05.461686Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 662, "message_tokens": 567, "message_utf8_bytes": 1429, "cache_tensor_bytes": null, "serialized_payload_bytes": 1429, "side_information_bytes": 0, "fetched_bytes": 809, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11846.203958033584, "ttft_ms": 13.124208897352219, "decode_ms": 13.026083004660904, "receiver_prefill_ms": 75.09020809084177}, "total_wall_ms": 11948.756666039117, "ttft_ms": 13.124208897352219, "input_usage": {"lookups": 3}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S1"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "8f98f10ad14c4a28016ef3eebb001913df39a6fa6ad55da93f968d1cd1f1eedf", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "9cd3561da56cbce8b341287f", "attempt_id": "d40e9db75672", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:05:06.389855Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 229072896}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 44, "message_tokens": null, "message_utf8_bytes": 30042, "cache_tensor_bytes": 229072896, "serialized_payload_bytes": 229072896, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 808.2617080071941, "ttft_ms": 52.685583010315895, "decode_ms": 136.16691704373807, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1006.2965829856694, "ttft_ms": 52.685583010315895, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4,644,474", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S1"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "45891fbe62951d2409b9d0682483fdbfd19a6f82c6a601a0f417aeaa6e93dd7f"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a57fa9dbd4dc6826a18ca305", "attempt_id": "924dcbe89cb1", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:05:06.606425Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 229072896}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 46, "message_tokens": null, "message_utf8_bytes": 30042, "cache_tensor_bytes": 229072896, "serialized_payload_bytes": 229072896, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 808.2617080071941, "ttft_ms": 46.51479097083211, "decode_ms": 149.76383408065885, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1011.7175000486895, "ttft_ms": 46.51479097083211, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "LP-81D-5209", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S1"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "94389922fe2c234a389aa721c8bb47d4597c8effef25316e72c62dd4a1aeecb7"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "4c54480affa3b17a5bb3b1ab", "attempt_id": "698045da49d9", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:05:06.827285Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 229072896}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 52, "message_tokens": null, "message_utf8_bytes": 30042, "cache_tensor_bytes": 229072896, "serialized_payload_bytes": 229072896, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 808.2617080071941, "ttft_ms": 48.611292033456266, "decode_ms": 105.46500002965331, "receiver_prefill_ms": 0.0}, "total_wall_ms": 968.8319999258965, "ttft_ms": 48.611292033456266, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5403805", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S1"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "ebd8c202e450e10a6d90509f598803c77ba6ed9c928c98df1eaddfa99eb333a4"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "17ca08da84df00e2336818c0", "attempt_id": "bdb9f6b09e20", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:05:07.005129Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 229072896}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 45, "message_tokens": null, "message_utf8_bytes": 30042, "cache_tensor_bytes": 229072896, "serialized_payload_bytes": 229072896, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 808.2617080071941, "ttft_ms": 49.74437493365258, "decode_ms": 104.32670905720443, "receiver_prefill_ms": 0.0}, "total_wall_ms": 969.8397080646828, "ttft_ms": 49.74437493365258, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5403805", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S1"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "ebd8c202e450e10a6d90509f598803c77ba6ed9c928c98df1eaddfa99eb333a4"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "254dbca5e6b7219a4eee2377", "attempt_id": "a885a26c5b83", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:05:07.186307Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 229072896}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 55, "message_tokens": null, "message_utf8_bytes": 30042, "cache_tensor_bytes": 229072896, "serialized_payload_bytes": 229072896, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 808.2617080071941, "ttft_ms": 47.35433298628777, "decode_ms": 89.7409999743104, "receiver_prefill_ms": 0.0}, "total_wall_ms": 952.8517499566078, "ttft_ms": 47.35433298628777, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Lithixymi-6", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S1"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "5919268fa866f12eb8b0497583904be835032b475afd544dc2eba31f2b4f6679"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "070988cf2c65dfe730f3734f", "attempt_id": "55ccbddeb80a", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:05:07.348552Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 229072896}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 30042, "cache_tensor_bytes": 229072896, "serialized_payload_bytes": 229072896, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 808.2617080071941, "ttft_ms": 48.46683400683105, "decode_ms": 148.71079195290804, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1012.6327080652118, "ttft_ms": 48.46683400683105, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "18,854,477", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S1"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "957852e0b5d7c6e6a90b05e9044fed309e018aed01ae6578b9dec8837c40895b"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "2fe045593632395888cbbd68", "attempt_id": "0a7c27a37875", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:05:07.571092Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 229072896}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 50, "message_tokens": null, "message_utf8_bytes": 30042, "cache_tensor_bytes": 229072896, "serialized_payload_bytes": 229072896, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 808.2617080071941, "ttft_ms": 45.76050001196563, "decode_ms": 15.173666994087398, "receiver_prefill_ms": 0.0}, "total_wall_ms": 876.8621659837663, "ttft_ms": 45.76050001196563, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S1"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "19d5676d424f819c306b1b37", "attempt_id": "a32d9c31e433", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:05:07.657004Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 229072896}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 49, "message_tokens": null, "message_utf8_bytes": 30042, "cache_tensor_bytes": 229072896, "serialized_payload_bytes": 229072896, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 808.2617080071941, "ttft_ms": 50.03195803146809, "decode_ms": 15.280457912012935, "receiver_prefill_ms": 0.0}, "total_wall_ms": 881.8116249749437, "ttft_ms": 50.03195803146809, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S1"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "0a8da2f5148e861cb3cd0e2b", "attempt_id": "3096ed23bf04", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:05:07.748469Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 6336, "message_tokens": 6250, "message_utf8_bytes": 30572, "cache_tensor_bytes": null, "serialized_payload_bytes": 30572, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.03287498839199543, "ttft_ms": 13.966750004328787, "decode_ms": 126.50008301716298, "receiver_prefill_ms": 786.3431660225615}, "total_wall_ms": 941.2742499262094, "ttft_ms": 13.966750004328787, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4,150,875", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S3"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "5867e0ad70e1bb5e0a752da28b19b936a361bc505de3b7b5b6e5215c8fc93c59", "output_hash": "44d2bea5683fa0161cfd0c3e09190160f393c874a2e7f4333b9ad36538effe28"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "17aedaa16bea2fee1ae4fd8b", "attempt_id": "70e242b69cae", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:05:08.707555Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 6338, "message_tokens": 6250, "message_utf8_bytes": 30572, "cache_tensor_bytes": null, "serialized_payload_bytes": 30572, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.03287498839199543, "ttft_ms": 14.252875000238419, "decode_ms": 139.03670804575086, "receiver_prefill_ms": 824.2297499673441}, "total_wall_ms": 987.3585830209777, "ttft_ms": 14.252875000238419, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "BY-78E-8447", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S3"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "5867e0ad70e1bb5e0a752da28b19b936a361bc505de3b7b5b6e5215c8fc93c59", "output_hash": "6acdf1f1807eeb525a30ee2d553ebf18ba3bb86ebe5cef8081992457efd3bb8d"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "fea02b51ac52bf307d822f18", "attempt_id": "954e1ddc6530", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:05:09.713430Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 6344, "message_tokens": 6250, "message_utf8_bytes": 30572, "cache_tensor_bytes": null, "serialized_payload_bytes": 30572, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.03287498839199543, "ttft_ms": 14.152207993902266, "decode_ms": 97.36274997703731, "receiver_prefill_ms": 845.2375839697197}, "total_wall_ms": 966.7567909928039, "ttft_ms": 14.152207993902266, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5403805", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S3"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "5867e0ad70e1bb5e0a752da28b19b936a361bc505de3b7b5b6e5215c8fc93c59", "output_hash": "ebd8c202e450e10a6d90509f598803c77ba6ed9c928c98df1eaddfa99eb333a4"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "30bdbe3e3d5cb59f77c46e33", "attempt_id": "e929f232d42e", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:05:10.698869Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 6337, "message_tokens": 6250, "message_utf8_bytes": 30572, "cache_tensor_bytes": null, "serialized_payload_bytes": 30572, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.03287498839199543, "ttft_ms": 14.046166092157364, "decode_ms": 96.11633303575218, "receiver_prefill_ms": 847.8993329918012}, "total_wall_ms": 967.933667008765, "ttft_ms": 14.046166092157364, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4150875", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S3"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "5867e0ad70e1bb5e0a752da28b19b936a361bc505de3b7b5b6e5215c8fc93c59", "output_hash": "9f908cf585946283f90239132608e9f36a14f3587bfad3942073463dff2947a8"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "bbf442ca9a91fe5691200078", "attempt_id": "bba5f64ee512", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:05:11.684113Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 6347, "message_tokens": 6250, "message_utf8_bytes": 30572, "cache_tensor_bytes": null, "serialized_payload_bytes": 30572, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.03287498839199543, "ttft_ms": 14.167250017635524, "decode_ms": 83.43433297704905, "receiver_prefill_ms": 839.8187080165371}, "total_wall_ms": 947.1073329914361, "ttft_ms": 14.167250017635524, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Lithixymi-6", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S3"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "5867e0ad70e1bb5e0a752da28b19b936a361bc505de3b7b5b6e5215c8fc93c59", "output_hash": "5919268fa866f12eb8b0497583904be835032b475afd544dc2eba31f2b4f6679"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "bb8cf03c94db6333dfff4164", "attempt_id": "b3f3c80e9dd1", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:05:12.648542Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 6339, "message_tokens": 6250, "message_utf8_bytes": 30572, "cache_tensor_bytes": null, "serialized_payload_bytes": 30572, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.03287498839199543, "ttft_ms": 14.354833052493632, "decode_ms": 99.05154106672853, "receiver_prefill_ms": 841.0300000105053}, "total_wall_ms": 964.0578329563141, "ttft_ms": 14.354833052493632, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "2318457", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S3"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "5867e0ad70e1bb5e0a752da28b19b936a361bc505de3b7b5b6e5215c8fc93c59", "output_hash": "97999c7916d0b7d84c47f51bc90c3717329cf46db25af686851cf2b11b856e5e"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "67a54457de449887316e67c6", "attempt_id": "2e80d88322b0", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:05:13.630327Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 6342, "message_tokens": 6250, "message_utf8_bytes": 30572, "cache_tensor_bytes": null, "serialized_payload_bytes": 30572, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.03287498839199543, "ttft_ms": 13.847875059582293, "decode_ms": 13.77433305606246, "receiver_prefill_ms": 834.2350410530344}, "total_wall_ms": 871.7559169745073, "ttft_ms": 13.847875059582293, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S3"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "5867e0ad70e1bb5e0a752da28b19b936a361bc505de3b7b5b6e5215c8fc93c59", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a91968683fc1476ea46390c7", "attempt_id": "84665b67997d", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:05:14.519465Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 6341, "message_tokens": 6250, "message_utf8_bytes": 30572, "cache_tensor_bytes": null, "serialized_payload_bytes": 30572, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.03287498839199543, "ttft_ms": 13.977374997921288, "decode_ms": 13.832999975420535, "receiver_prefill_ms": 846.0721250157803}, "total_wall_ms": 883.3868330111727, "ttft_ms": 13.977374997921288, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S3"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "5867e0ad70e1bb5e0a752da28b19b936a361bc505de3b7b5b6e5215c8fc93c59", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "3edeef7c231f4e719397b224", "attempt_id": "2881340fea6d", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:05:41.477455Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1885}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 1999, "message_tokens": 1912, "message_utf8_bytes": 6070, "cache_tensor_bytes": null, "serialized_payload_bytes": 6070, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26056.91695807036, "ttft_ms": 13.430959079414606, "decode_ms": 120.57104194536805, "receiver_prefill_ms": 210.0348339881748}, "total_wall_ms": 26404.70295806881, "ttft_ms": 13.430959079414606, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4,644,474", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S3"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "fcec459a75578d55c7831a6378635f4f6be218bd4d3bd2a46d16962c2b153ea4", "output_hash": "45891fbe62951d2409b9d0682483fdbfd19a6f82c6a601a0f417aeaa6e93dd7f"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "3c4f10b8dfee6f18caf97acf", "attempt_id": "ab584a531fbd", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:05:41.843297Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1885}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 2001, "message_tokens": 1912, "message_utf8_bytes": 6070, "cache_tensor_bytes": null, "serialized_payload_bytes": 6070, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26056.91695807036, "ttft_ms": 13.413458014838398, "decode_ms": 132.93933391105384, "receiver_prefill_ms": 208.54970905929804}, "total_wall_ms": 26414.784708060324, "ttft_ms": 13.413458014838398, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "BY-78E-8447", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S3"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "fcec459a75578d55c7831a6378635f4f6be218bd4d3bd2a46d16962c2b153ea4", "output_hash": "6acdf1f1807eeb525a30ee2d553ebf18ba3bb86ebe5cef8081992457efd3bb8d"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b7193590ac412436ca29a85b", "attempt_id": "f5c6d499b729", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:05:42.218706Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1885}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 2007, "message_tokens": 1912, "message_utf8_bytes": 6070, "cache_tensor_bytes": null, "serialized_payload_bytes": 6070, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26056.91695807036, "ttft_ms": 13.672166038304567, "decode_ms": 94.75333301816136, "receiver_prefill_ms": 212.8239170415327}, "total_wall_ms": 26381.15149899386, "ttft_ms": 13.672166038304567, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5403805", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S3"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "fcec459a75578d55c7831a6378635f4f6be218bd4d3bd2a46d16962c2b153ea4", "output_hash": "ebd8c202e450e10a6d90509f598803c77ba6ed9c928c98df1eaddfa99eb333a4"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "cb78ba93f4b0f59db719e1a9", "attempt_id": "57e2c742b878", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:05:42.562853Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1885}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 2000, "message_tokens": 1912, "message_utf8_bytes": 6070, "cache_tensor_bytes": null, "serialized_payload_bytes": 6070, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26056.91695807036, "ttft_ms": 14.372666948474944, "decode_ms": 95.2110419748351, "receiver_prefill_ms": 237.8027499653399}, "total_wall_ms": 26407.28712512646, "ttft_ms": 14.372666948474944, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5403805", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S3"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "fcec459a75578d55c7831a6378635f4f6be218bd4d3bd2a46d16962c2b153ea4", "output_hash": "ebd8c202e450e10a6d90509f598803c77ba6ed9c928c98df1eaddfa99eb333a4"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "7894fce7ca138fcdb845820c", "attempt_id": "514694f98e26", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:05:42.930844Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1885}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 2010, "message_tokens": 1912, "message_utf8_bytes": 6070, "cache_tensor_bytes": null, "serialized_payload_bytes": 6070, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26056.91695807036, "ttft_ms": 13.758916989900172, "decode_ms": 94.18604196980596, "receiver_prefill_ms": 238.50908409804106}, "total_wall_ms": 26406.3592919847, "ttft_ms": 13.758916989900172, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Keryminov-10", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S3"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "fcec459a75578d55c7831a6378635f4f6be218bd4d3bd2a46d16962c2b153ea4", "output_hash": "47d4fb882fdd2db10f1e88942b3f981e24a10a4a4501ff56cdaf321c9eba1c18"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "9f1406c94af84e08f138257d", "attempt_id": "ae30bbb55c6a", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:05:43.297614Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1885}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 2002, "message_tokens": 1912, "message_utf8_bytes": 6070, "cache_tensor_bytes": null, "serialized_payload_bytes": 6070, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26056.91695807036, "ttft_ms": 13.745333068072796, "decode_ms": 96.32583300117403, "receiver_prefill_ms": 232.71737492177635}, "total_wall_ms": 26402.777833049186, "ttft_ms": 13.745333068072796, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7653745", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S3"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "fcec459a75578d55c7831a6378635f4f6be218bd4d3bd2a46d16962c2b153ea4", "output_hash": "ffac6cdafbd3ff7cc776cf8c458a4e330b7a4a2d9015c5d9abe4555d930c52e4"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "fb88906ffcd33ad6140bf840", "attempt_id": "037c78c69cce", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:05:43.660673Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1885}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 2005, "message_tokens": 1912, "message_utf8_bytes": 6070, "cache_tensor_bytes": null, "serialized_payload_bytes": 6070, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26056.91695807036, "ttft_ms": 13.76820809673518, "decode_ms": 13.563416083343327, "receiver_prefill_ms": 233.4959589643404}, "total_wall_ms": 26320.709875086322, "ttft_ms": 13.76820809673518, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S3"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "fcec459a75578d55c7831a6378635f4f6be218bd4d3bd2a46d16962c2b153ea4", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "1c22b26471e65af94b6b7f94", "attempt_id": "ca068cc6ca7f", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:05:43.942500Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1885}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 2004, "message_tokens": 1912, "message_utf8_bytes": 6070, "cache_tensor_bytes": null, "serialized_payload_bytes": 6070, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26056.91695807036, "ttft_ms": 13.804375077597797, "decode_ms": 13.66379193495959, "receiver_prefill_ms": 239.7810829570517}, "total_wall_ms": 26327.03562406823, "ttft_ms": 13.804375077597797, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S3"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "fcec459a75578d55c7831a6378635f4f6be218bd4d3bd2a46d16962c2b153ea4", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "e9df0ffc9e6183082a79a026", "attempt_id": "4e8f003538d5", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:06:10.090444Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 28}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 1208, "message_tokens": 1117, "message_utf8_bytes": 2984, "cache_tensor_bytes": null, "serialized_payload_bytes": 2984, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25860.254374914803, "ttft_ms": 13.639541924931109, "decode_ms": 13.261166983284056, "receiver_prefill_ms": 128.0348339350894}, "total_wall_ms": 26022.08508295007, "ttft_ms": 13.639541924931109, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S3"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "d77463e358c32a351350b28bb0cc8a08eb73c949f731c488dbac2fb96706a60f", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "121dcd83f2504e2b88f02dc5", "attempt_id": "69b369abd4a6", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:06:10.271018Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 28}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 1210, "message_tokens": 1117, "message_utf8_bytes": 2984, "cache_tensor_bytes": null, "serialized_payload_bytes": 2984, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25860.254374914803, "ttft_ms": 13.25616694521159, "decode_ms": 13.147082994692028, "receiver_prefill_ms": 126.28491595387459}, "total_wall_ms": 26014.798416988924, "ttft_ms": 13.25616694521159, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S3"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "d77463e358c32a351350b28bb0cc8a08eb73c949f731c488dbac2fb96706a60f", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "5fecf9a61966efcbe0178937", "attempt_id": "174a6660a5fe", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:06:10.442754Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 28}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 1216, "message_tokens": 1117, "message_utf8_bytes": 2984, "cache_tensor_bytes": null, "serialized_payload_bytes": 2984, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25860.254374914803, "ttft_ms": 14.118375023826957, "decode_ms": 97.91758295614272, "receiver_prefill_ms": 431.37695803306997}, "total_wall_ms": 26405.521207954735, "ttft_ms": 14.118375023826957, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4150875", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S3"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "d77463e358c32a351350b28bb0cc8a08eb73c949f731c488dbac2fb96706a60f", "output_hash": "9f908cf585946283f90239132608e9f36a14f3587bfad3942073463dff2947a8"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a19cc3c04bdbc22d08698f29", "attempt_id": "dc240261f370", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:06:11.006006Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 28}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 1209, "message_tokens": 1117, "message_utf8_bytes": 2984, "cache_tensor_bytes": null, "serialized_payload_bytes": 2984, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25860.254374914803, "ttft_ms": 14.133582939393818, "decode_ms": 99.780207965523, "receiver_prefill_ms": 532.4197920272127}, "total_wall_ms": 26508.506541955285, "ttft_ms": 14.133582939393818, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4150875", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S3"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "d77463e358c32a351350b28bb0cc8a08eb73c949f731c488dbac2fb96706a60f", "output_hash": "9f908cf585946283f90239132608e9f36a14f3587bfad3942073463dff2947a8"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "20548acf102200b6060bdf57", "attempt_id": "98070833a751", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:06:11.673365Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 28}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 1219, "message_tokens": 1117, "message_utf8_bytes": 2984, "cache_tensor_bytes": null, "serialized_payload_bytes": 2984, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25860.254374914803, "ttft_ms": 13.808291987515986, "decode_ms": 82.57945801597089, "receiver_prefill_ms": 705.7896659243852}, "total_wall_ms": 26664.71245896537, "ttft_ms": 13.808291987515986, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Lithixymi-6", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S3"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "d77463e358c32a351350b28bb0cc8a08eb73c949f731c488dbac2fb96706a60f", "output_hash": "5919268fa866f12eb8b0497583904be835032b475afd544dc2eba31f2b4f6679"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "d1ec3a8807a73fe4cbd0d6e7", "attempt_id": "dde5ef4d7e84", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:06:12.496965Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 28}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 1211, "message_tokens": 1117, "message_utf8_bytes": 2984, "cache_tensor_bytes": null, "serialized_payload_bytes": 2984, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25860.254374914803, "ttft_ms": 13.21308291517198, "decode_ms": 94.33804196305573, "receiver_prefill_ms": 165.7823328860104}, "total_wall_ms": 26135.553957894444, "ttft_ms": 13.21308291517198, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1455908", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S3"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "d77463e358c32a351350b28bb0cc8a08eb73c949f731c488dbac2fb96706a60f", "output_hash": "2fa9e0be664f4b3bf1181594e5273b979370140586a6b43009e76be33070b50c"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "9bdee3d1224407ba608430ce", "attempt_id": "058d9284c579", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:06:12.797837Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 28}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 1214, "message_tokens": 1117, "message_utf8_bytes": 2984, "cache_tensor_bytes": null, "serialized_payload_bytes": 2984, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25860.254374914803, "ttft_ms": 13.314165989868343, "decode_ms": 13.583249994553626, "receiver_prefill_ms": 141.8073340319097}, "total_wall_ms": 26030.847791931592, "ttft_ms": 13.314165989868343, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S3"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "d77463e358c32a351350b28bb0cc8a08eb73c949f731c488dbac2fb96706a60f", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "03f0c5bf7cbf2aa0cc80904d", "attempt_id": "7e91b75c517f", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:06:12.985880Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 28}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 1213, "message_tokens": 1117, "message_utf8_bytes": 2984, "cache_tensor_bytes": null, "serialized_payload_bytes": 2984, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25860.254374914803, "ttft_ms": 13.224207912571728, "decode_ms": 13.414583052508533, "receiver_prefill_ms": 126.4943330315873}, "total_wall_ms": 26015.22674993612, "ttft_ms": 13.224207912571728, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S3"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "d77463e358c32a351350b28bb0cc8a08eb73c949f731c488dbac2fb96706a60f", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "cab755331822333cf3184c3e", "attempt_id": "497b76e80f2b", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:06:39.080587Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 2285, "message_tokens": 2195, "message_utf8_bytes": 5743, "cache_tensor_bytes": null, "serialized_payload_bytes": 5743, "side_information_bytes": 0, "fetched_bytes": 2895, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25921.86312493868, "ttft_ms": 13.354249997064471, "decode_ms": 13.265292043797672, "receiver_prefill_ms": 249.4450839003548}, "total_wall_ms": 26202.304832986556, "ttft_ms": 13.354249997064471, "input_usage": {"lookups": 10}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S3"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a2852e152894649046714a3da69688a401c739f4eb401a22d71a98e45e7de610", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "4625d1642770d31fc3a4e259", "attempt_id": "f7568f6ae014", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:06:39.380500Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 2164, "message_tokens": 2072, "message_utf8_bytes": 5485, "cache_tensor_bytes": null, "serialized_payload_bytes": 5485, "side_information_bytes": 0, "fetched_bytes": 2631, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25921.86312493868, "ttft_ms": 13.365708058699965, "decode_ms": 134.0603749267757, "receiver_prefill_ms": 240.85054208990186}, "total_wall_ms": 26314.05924994033, "ttft_ms": 13.365708058699965, "input_usage": {"lookups": 10}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "MS-90S-7041", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S3"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "2fb364c9d1cc847b1ace32764d3504ac86f9f3d532a1d25892181a99c81672a6", "output_hash": "1e3d38ee94090be1d82d2ec022cd718f1c7c52aaeed713007a50cca23e4215a9"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "0972f2afc2a23dee202ac5b3", "attempt_id": "747092386ce3", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:06:39.790027Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 2276, "message_tokens": 2178, "message_utf8_bytes": 5702, "cache_tensor_bytes": null, "serialized_payload_bytes": 5702, "side_information_bytes": 0, "fetched_bytes": 2859, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25921.86312493868, "ttft_ms": 13.721500057727098, "decode_ms": 95.836375025101, "receiver_prefill_ms": 258.7239999556914}, "total_wall_ms": 26293.304833001457, "ttft_ms": 13.721500057727098, "input_usage": {"lookups": 10}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5403805", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S3"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "4ab9eb21ed79c88b53bfdad9067bff1387604993542ee850c0ec9934ba3f3681", "output_hash": "ebd8c202e450e10a6d90509f598803c77ba6ed9c928c98df1eaddfa99eb333a4"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "5512b0816f7c340935715755", "attempt_id": "1b95534b0c2c", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:06:40.179266Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 2269, "message_tokens": 2178, "message_utf8_bytes": 5702, "cache_tensor_bytes": null, "serialized_payload_bytes": 5702, "side_information_bytes": 0, "fetched_bytes": 2859, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25921.86312493868, "ttft_ms": 13.894791016355157, "decode_ms": 94.73695792257786, "receiver_prefill_ms": 305.0927920266986}, "total_wall_ms": 26338.66758295335, "ttft_ms": 13.894791016355157, "input_usage": {"lookups": 10}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4150875", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S3"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "4ab9eb21ed79c88b53bfdad9067bff1387604993542ee850c0ec9934ba3f3681", "output_hash": "9f908cf585946283f90239132608e9f36a14f3587bfad3942073463dff2947a8"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "1cd7922885c308a5f53215ec", "attempt_id": "df54d3f90d0b", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:06:40.614167Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 2217, "message_tokens": 2116, "message_utf8_bytes": 5543, "cache_tensor_bytes": null, "serialized_payload_bytes": 5543, "side_information_bytes": 0, "fetched_bytes": 2725, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25921.86312493868, "ttft_ms": 13.766208081506193, "decode_ms": 80.77875000890344, "receiver_prefill_ms": 276.7068329267204}, "total_wall_ms": 26296.097374986857, "ttft_ms": 13.766208081506193, "input_usage": {"lookups": 10}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Lithixymi-6", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S3"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "daf9cbcb1be7297fe328a3efdacaccff678f6ac18dd6676c707ac3f376e3d1df", "output_hash": "5919268fa866f12eb8b0497583904be835032b475afd544dc2eba31f2b4f6679"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c5ff71cfd9ff778d3d55ab5c", "attempt_id": "60f7f7fd7619", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:06:41.005679Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 2288, "message_tokens": 2195, "message_utf8_bytes": 5743, "cache_tensor_bytes": null, "serialized_payload_bytes": 5743, "side_information_bytes": 0, "fetched_bytes": 2895, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25921.86312493868, "ttft_ms": 13.603665982373059, "decode_ms": 135.28054195921868, "receiver_prefill_ms": 270.8447080804035}, "total_wall_ms": 26344.598582945764, "ttft_ms": 13.603665982373059, "input_usage": {"lookups": 10}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "18,339,433", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S3"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "0bce4ff4cde80c1459561b67907d44c30baef4771161f98038b3e3791ad26c0f", "output_hash": "64b4b1dc64d8c623ecdd5ad06ce9db616a306fb205f03da7c9577dd3000ec66a"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "e7a618dd6d86c7510cf3f9bc", "attempt_id": "7024f69198f6", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:06:41.446328Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 2218, "message_tokens": 2122, "message_utf8_bytes": 5682, "cache_tensor_bytes": null, "serialized_payload_bytes": 5682, "side_information_bytes": 0, "fetched_bytes": 2807, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25921.86312493868, "ttft_ms": 13.583665946498513, "decode_ms": 13.713416992686689, "receiver_prefill_ms": 264.3962079891935}, "total_wall_ms": 26216.528541990556, "ttft_ms": 13.583665946498513, "input_usage": {"lookups": 10}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S3"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "844371e8e15a55b93443e383a4de2ccd133778198574f0c5cb6c0c8f127ddae1", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "8d6093c8f1829e2b4299b72d", "attempt_id": "8f891c11fea1", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:06:41.758525Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 2290, "message_tokens": 2195, "message_utf8_bytes": 5743, "cache_tensor_bytes": null, "serialized_payload_bytes": 5743, "side_information_bytes": 0, "fetched_bytes": 2895, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25921.86312493868, "ttft_ms": 13.95733398385346, "decode_ms": 13.765500043518841, "receiver_prefill_ms": 278.7422079127282}, "total_wall_ms": 26231.380582903512, "ttft_ms": 13.95733398385346, "input_usage": {"lookups": 10}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S3"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "6aeb3f8a7f547b7d0eda1a79fbd703245b35a5c46afdac9df0df8dda2de526e0", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "9d414bf048cb43a45e327179", "attempt_id": "347b509c17f2", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:06:43.119126Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 236150784}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 44, "message_tokens": null, "message_utf8_bytes": 30518, "cache_tensor_bytes": 236150784, "serialized_payload_bytes": 236150784, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 1033.0043339636177, "ttft_ms": 54.75395894609392, "decode_ms": 108.54862502310425, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1237.2092918958515, "ttft_ms": 54.75395894609392, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "464,470", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S3"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "835de2f00202e4eec0327b4fbf8c5d958bf8be1bcbd34ce647640374b61fd2bd"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "4854b19d575fb743f0545c32", "attempt_id": "852ab7a0732c", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:06:43.351418Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 236150784}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 46, "message_tokens": null, "message_utf8_bytes": 30518, "cache_tensor_bytes": 236150784, "serialized_payload_bytes": 236150784, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 1033.0043339636177, "ttft_ms": 47.00400005094707, "decode_ms": 155.3390830522403, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1239.715668023564, "ttft_ms": 47.00400005094707, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "UU-99A-2257", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S3"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "05a311075f0ae9abae2c84c2f3eaf1f9ee9d38e801a7a116005be8bd9d71a8fe"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "d84a9e0bc396ee572299e604", "attempt_id": "79bfaf772ef6", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:06:43.585804Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 236150784}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 52, "message_tokens": null, "message_utf8_bytes": 30518, "cache_tensor_bytes": 236150784, "serialized_payload_bytes": 236150784, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 1033.0043339636177, "ttft_ms": 182.67491599544883, "decode_ms": 109.15449995081872, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1329.2211249936372, "ttft_ms": 182.67491599544883, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "292,875", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S3"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "ff4619747cebcce000996a16d9dcda711eb004061946a6e41ecbcfe01a83832b"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "9edb9cf27c4be27c3eced52e", "attempt_id": "126d63117b50", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:06:43.900666Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 236150784}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 45, "message_tokens": null, "message_utf8_bytes": 30518, "cache_tensor_bytes": 236150784, "serialized_payload_bytes": 236150784, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 1033.0043339636177, "ttft_ms": 84.29658296518028, "decode_ms": 104.57575006876141, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1226.2898759217933, "ttft_ms": 84.29658296518028, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "335,480", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S3"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "024b89cda0bec61fddd92838da3e87403abc1572260f74656cd1f0bffcfc2434"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "7fea884881b9de7c013a1c14", "attempt_id": "15d52c87b3df", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:06:44.111969Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 236150784}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 55, "message_tokens": null, "message_utf8_bytes": 30518, "cache_tensor_bytes": 236150784, "serialized_payload_bytes": 236150784, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 1033.0043339636177, "ttft_ms": 48.28091699164361, "decode_ms": 109.34591700788587, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1195.1411250047386, "ttft_ms": 48.28091699164361, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Keryminov-10", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S3"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "47d4fb882fdd2db10f1e88942b3f981e24a10a4a4501ff56cdaf321c9eba1c18"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c93e050f700c9f8d90fd2fc7", "attempt_id": "27f29d3e1209", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:06:44.291375Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 236150784}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 30518, "cache_tensor_bytes": 236150784, "serialized_payload_bytes": 236150784, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 1033.0043339636177, "ttft_ms": 49.073707894422114, "decode_ms": 119.7517920518294, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1205.4830839624628, "ttft_ms": 49.073707894422114, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "14229228", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S3"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "24732933aa145f48f1a25eb6b64227af90537fbab4e761a8d0d07e1d2401568b"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "123fb05a26953bcf76e8d5d0", "attempt_id": "207704682dd8", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:06:44.481732Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 236150784}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 50, "message_tokens": null, "message_utf8_bytes": 30518, "cache_tensor_bytes": 236150784, "serialized_payload_bytes": 236150784, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 1033.0043339636177, "ttft_ms": 46.9366250326857, "decode_ms": 14.956166967749596, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1099.5982930762693, "ttft_ms": 46.9366250326857, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S3"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "2f7c1e2230da2e7fbcd794f1", "attempt_id": "6790ecf0ce25", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:06:44.566365Z", "task_family": "synthetic", "document_id": "syn_03007_8000", "document_hash": "601a34ce3fb4eb4fe6346a3bc9900160762a432b30f192472ef5821aa5ad3f6f", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 236150784}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5843, "query_tokens": 49, "message_tokens": null, "message_utf8_bytes": 30518, "cache_tensor_bytes": 236150784, "serialized_payload_bytes": 236150784, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 1033.0043339636177, "ttft_ms": 50.98741699475795, "decode_ms": 15.142875025048852, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1103.4820009954274, "ttft_ms": 50.98741699475795, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Eshard-1", "bogus_amount": "6,004,335", "source": "S3"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03007_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b283db186d7c035874ad58a5", "attempt_id": "95f857e31661", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:06:45.575752Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 2369, "message_tokens": 2283, "message_utf8_bytes": 8329, "cache_tensor_bytes": null, "serialized_payload_bytes": 8329, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.02766598481684923, "ttft_ms": 13.515333994291723, "decode_ms": 93.25137489940971, "receiver_prefill_ms": 270.6569581059739}, "total_wall_ms": 381.4585000509396, "ttft_ms": 13.515333994291723, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4670383", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S1"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "98ab796b8615baaca2dc4b3a76d9a5c030e1469b387170d36fabd44f1aec6c05", "output_hash": "19ae8c357e5500bf494a6163bfbd11380b2996e8ea532a2c64f4a599464820f7"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "4d42349aa48760e1656be3db", "attempt_id": "2d36e5cd7f3b", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:06:45.975671Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 2372, "message_tokens": 2283, "message_utf8_bytes": 8329, "cache_tensor_bytes": null, "serialized_payload_bytes": 8329, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.02766598481684923, "ttft_ms": 13.531708042137325, "decode_ms": 134.00425005238503, "receiver_prefill_ms": 261.3682079827413}, "total_wall_ms": 412.83874900545925, "ttft_ms": 13.531708042137325, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "NA-58Q-5913", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S1"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "98ab796b8615baaca2dc4b3a76d9a5c030e1469b387170d36fabd44f1aec6c05", "output_hash": "a4952c0aa4af7f7a21616de82a1eb2bd97ee62e7fa53ca310442860a9bbc6d44"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "ef71dd550be7efd2f8d9b5cb", "attempt_id": "dd991e5ac46c", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:06:46.402252Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 2379, "message_tokens": 2283, "message_utf8_bytes": 8329, "cache_tensor_bytes": null, "serialized_payload_bytes": 8329, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.02766598481684923, "ttft_ms": 13.764959061518312, "decode_ms": 107.49045794364065, "receiver_prefill_ms": 261.9537499267608}, "total_wall_ms": 386.81745808571577, "ttft_ms": 13.764959061518312, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "10047105", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S1"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "98ab796b8615baaca2dc4b3a76d9a5c030e1469b387170d36fabd44f1aec6c05", "output_hash": "bea99610d1e3e75a7f20939faf8e6824e02237b156dacd6442ba2a7e02617510"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "1b04c7498e62a5b079c34c63", "attempt_id": "21eae45836fd", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:06:46.803532Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 2372, "message_tokens": 2283, "message_utf8_bytes": 8329, "cache_tensor_bytes": null, "serialized_payload_bytes": 8329, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.02766598481684923, "ttft_ms": 13.545499998144805, "decode_ms": 107.47466702014208, "receiver_prefill_ms": 261.50470797438174}, "total_wall_ms": 386.1227079760283, "ttft_ms": 13.545499998144805, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "10047105", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S1"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "98ab796b8615baaca2dc4b3a76d9a5c030e1469b387170d36fabd44f1aec6c05", "output_hash": "bea99610d1e3e75a7f20939faf8e6824e02237b156dacd6442ba2a7e02617510"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "0bd4955ceef46597701e2437", "attempt_id": "b7a0002504ff", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:06:47.203789Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 2381, "message_tokens": 2283, "message_utf8_bytes": 8329, "cache_tensor_bytes": null, "serialized_payload_bytes": 8329, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.02766598481684923, "ttft_ms": 13.849709066562355, "decode_ms": 93.50445901509374, "receiver_prefill_ms": 261.96466700639576}, "total_wall_ms": 372.91491602081805, "ttft_ms": 13.849709066562355, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Koluntek-19", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S1"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "98ab796b8615baaca2dc4b3a76d9a5c030e1469b387170d36fabd44f1aec6c05", "output_hash": "cefa71c3350f53880f3dfdde8949de2182d8e5275946ac0f441e8068e4a6282a"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "ac2dc0956fd15022cb652471", "attempt_id": "fac5a4eaa875", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:06:47.591682Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 2372, "message_tokens": 2283, "message_utf8_bytes": 8329, "cache_tensor_bytes": null, "serialized_payload_bytes": 8329, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.02766598481684923, "ttft_ms": 13.506041024811566, "decode_ms": 93.46312494017184, "receiver_prefill_ms": 260.93762496020645}, "total_wall_ms": 371.42612400930375, "ttft_ms": 13.506041024811566, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "2106495", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S1"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "98ab796b8615baaca2dc4b3a76d9a5c030e1469b387170d36fabd44f1aec6c05", "output_hash": "25aa9d44fe3cf34a5ca727765cd50092fb4af2999d15b8736d10a35d6c33a0b5"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "e9765635fb8099239900a96d", "attempt_id": "d27e654eb5ac", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:06:47.977458Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 2374, "message_tokens": 2283, "message_utf8_bytes": 8329, "cache_tensor_bytes": null, "serialized_payload_bytes": 8329, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.02766598481684923, "ttft_ms": 13.431166997179389, "decode_ms": 13.535959064029157, "receiver_prefill_ms": 261.45799993537366}, "total_wall_ms": 292.05379099585116, "ttft_ms": 13.431166997179389, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S1"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "98ab796b8615baaca2dc4b3a76d9a5c030e1469b387170d36fabd44f1aec6c05", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "bc5719406650d4dcd17160bd", "attempt_id": "3340f7fad87b", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:06:48.286711Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 2374, "message_tokens": 2283, "message_utf8_bytes": 8329, "cache_tensor_bytes": null, "serialized_payload_bytes": 8329, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.02766598481684923, "ttft_ms": 13.335624942556024, "decode_ms": 13.59733403660357, "receiver_prefill_ms": 261.36620808392763}, "total_wall_ms": 291.77958297077566, "ttft_ms": 13.335624942556024, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S1"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "98ab796b8615baaca2dc4b3a76d9a5c030e1469b387170d36fabd44f1aec6c05", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "47067043cdd3f394b579be5e", "attempt_id": "102474d6e138", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:07:02.551502Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 1125, "message_tokens": 1037, "message_utf8_bytes": 2705, "cache_tensor_bytes": null, "serialized_payload_bytes": 2705, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13959.18991707731, "ttft_ms": 13.379542040638626, "decode_ms": 121.01179198361933, "receiver_prefill_ms": 121.32512510288507}, "total_wall_ms": 14221.78095811978, "ttft_ms": 13.379542040638626, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4,670,383", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S1"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "f74e69e01012aa03dc28b9dd803a89d65efd9cf48eec47f8a173d84e2b1fbdc5", "output_hash": "9510633f2e349a1cedecda578083181521646f70ca5ae08bae4b9188f3bd4a3b"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "3f30d89fa4d0a3bb10ec7863", "attempt_id": "455e3702161a", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:07:02.828663Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 1128, "message_tokens": 1037, "message_utf8_bytes": 2705, "cache_tensor_bytes": null, "serialized_payload_bytes": 2705, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13959.18991707731, "ttft_ms": 13.308166991919279, "decode_ms": 133.62333294935524, "receiver_prefill_ms": 121.02620908990502}, "total_wall_ms": 14228.959709056653, "ttft_ms": 13.308166991919279, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "NA-58Q-5913", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S1"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "f74e69e01012aa03dc28b9dd803a89d65efd9cf48eec47f8a173d84e2b1fbdc5", "output_hash": "a4952c0aa4af7f7a21616de82a1eb2bd97ee62e7fa53ca310442860a9bbc6d44"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c228190d942a5088b77ee222", "attempt_id": "13b12e9c992c", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:07:03.112877Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 1135, "message_tokens": 1037, "message_utf8_bytes": 2705, "cache_tensor_bytes": null, "serialized_payload_bytes": 2705, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13959.18991707731, "ttft_ms": 13.427750091068447, "decode_ms": 92.84641698468477, "receiver_prefill_ms": 121.78983294870704}, "total_wall_ms": 14189.030417008325, "ttft_ms": 13.427750091068447, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "9354593", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S1"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "f74e69e01012aa03dc28b9dd803a89d65efd9cf48eec47f8a173d84e2b1fbdc5", "output_hash": "7f6c515b7ed67e83d3ff5b2f273117e392e33dc75882fe056469fb6321f3141c"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "cdacb6f610e7d0051a0504af", "attempt_id": "791598749206", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:07:03.357045Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 1128, "message_tokens": 1037, "message_utf8_bytes": 2705, "cache_tensor_bytes": null, "serialized_payload_bytes": 2705, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13959.18991707731, "ttft_ms": 13.154541957192123, "decode_ms": 92.71933394484222, "receiver_prefill_ms": 120.85987499449402}, "total_wall_ms": 14187.70554207731, "ttft_ms": 13.154541957192123, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "9354593", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S1"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "f74e69e01012aa03dc28b9dd803a89d65efd9cf48eec47f8a173d84e2b1fbdc5", "output_hash": "7f6c515b7ed67e83d3ff5b2f273117e392e33dc75882fe056469fb6321f3141c"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "538fe33c800f4c9a467f666d", "attempt_id": "d344a4276a76", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:07:03.600019Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 1137, "message_tokens": 1037, "message_utf8_bytes": 2705, "cache_tensor_bytes": null, "serialized_payload_bytes": 2705, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13959.18991707731, "ttft_ms": 13.54979199822992, "decode_ms": 92.9634579224512, "receiver_prefill_ms": 121.08091707341373}, "total_wall_ms": 14188.579167006537, "ttft_ms": 13.54979199822992, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Eshgamgam-11", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S1"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "f74e69e01012aa03dc28b9dd803a89d65efd9cf48eec47f8a173d84e2b1fbdc5", "output_hash": "8c1313c53ab4dbffa5f336d721c810678133eecaa467241e53d8a92150601475"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "1f973b08c4f4d4ce51334cf5", "attempt_id": "95b167d3b705", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:07:03.843472Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 1128, "message_tokens": 1037, "message_utf8_bytes": 2705, "cache_tensor_bytes": null, "serialized_payload_bytes": 2705, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13959.18991707731, "ttft_ms": 13.602083083242178, "decode_ms": 133.62745789345354, "receiver_prefill_ms": 121.11212499439716}, "total_wall_ms": 14229.412917047739, "ttft_ms": 13.602083083242178, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "15,432,677", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S1"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "f74e69e01012aa03dc28b9dd803a89d65efd9cf48eec47f8a173d84e2b1fbdc5", "output_hash": "69b46804e5a33ea1c99182d3fbd43960616c3d7b9e8907ea5222c5c5f5a511fa"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "efeb06c9d691ca28521980ec", "attempt_id": "4ad10b1aa0fb", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:07:04.128185Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 1130, "message_tokens": 1037, "message_utf8_bytes": 2705, "cache_tensor_bytes": null, "serialized_payload_bytes": 2705, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13959.18991707731, "ttft_ms": 13.171541970223188, "decode_ms": 13.23949999641627, "receiver_prefill_ms": 121.5072909835726}, "total_wall_ms": 14108.861667104065, "ttft_ms": 13.171541970223188, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S1"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "f74e69e01012aa03dc28b9dd803a89d65efd9cf48eec47f8a173d84e2b1fbdc5", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "ec1a1500098391e50214682f", "attempt_id": "21fc40e32de0", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:07:04.291905Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 1130, "message_tokens": 1037, "message_utf8_bytes": 2705, "cache_tensor_bytes": null, "serialized_payload_bytes": 2705, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13959.18991707731, "ttft_ms": 13.395000016316772, "decode_ms": 13.341124984435737, "receiver_prefill_ms": 123.69783292524517}, "total_wall_ms": 14111.345126060769, "ttft_ms": 13.395000016316772, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S1"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "f74e69e01012aa03dc28b9dd803a89d65efd9cf48eec47f8a173d84e2b1fbdc5", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "ef109c47c2a6f3cdaff8f6ec", "attempt_id": "b12a0cc65fee", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:07:12.396334Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 8}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 419, "message_tokens": 328, "message_utf8_bytes": 911, "cache_tensor_bytes": null, "serialized_payload_bytes": 911, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 7938.297540997155, "ttft_ms": 13.102666940540075, "decode_ms": 12.951915967278183, "receiver_prefill_ms": 53.65716607775539}, "total_wall_ms": 8019.440581905656, "ttft_ms": 13.102666940540075, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S1"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "c9648fa6d5cf3d28dcd311f0167cc512f617fded0e15a811bedfd58d5afd86b9", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "886b4559008e8b12635efb90", "attempt_id": "e13a2cfcbf87", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:07:12.497215Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 8}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 422, "message_tokens": 328, "message_utf8_bytes": 911, "cache_tensor_bytes": null, "serialized_payload_bytes": 911, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 7938.297540997155, "ttft_ms": 13.077791081741452, "decode_ms": 92.08550001494586, "receiver_prefill_ms": 53.54054202325642}, "total_wall_ms": 8098.081290954724, "ttft_ms": 13.077791081741452, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Project Wenvolqui-4", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S1"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "c9648fa6d5cf3d28dcd311f0167cc512f617fded0e15a811bedfd58d5afd86b9", "output_hash": "bb0ff91091f1013afd4d8296f8a2dd1cd0e54e2c53189d8c26af88494fecdaee"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "bcbac744b01be5a4e40e82fd", "attempt_id": "1fffa66c29b7", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:07:12.671580Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 8}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 429, "message_tokens": 328, "message_utf8_bytes": 911, "cache_tensor_bytes": null, "serialized_payload_bytes": 911, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 7938.297540997155, "ttft_ms": 13.34479101933539, "decode_ms": 13.040709076449275, "receiver_prefill_ms": 53.47037501633167}, "total_wall_ms": 8019.291082047857, "ttft_ms": 13.34479101933539, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S1"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "c9648fa6d5cf3d28dcd311f0167cc512f617fded0e15a811bedfd58d5afd86b9", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "9134ab49ce897a75f8d85578", "attempt_id": "ff204783c0ee", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:07:12.766819Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 8}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 422, "message_tokens": 328, "message_utf8_bytes": 911, "cache_tensor_bytes": null, "serialized_payload_bytes": 911, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 7938.297540997155, "ttft_ms": 13.236833037808537, "decode_ms": 13.337707961909473, "receiver_prefill_ms": 53.37787501048297}, "total_wall_ms": 8019.337290897965, "ttft_ms": 13.236833037808537, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S1"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "c9648fa6d5cf3d28dcd311f0167cc512f617fded0e15a811bedfd58d5afd86b9", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c35b440ef6d6b9e8783d6944", "attempt_id": "08bf9ea23fec", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:07:12.862415Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 8}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 431, "message_tokens": 328, "message_utf8_bytes": 911, "cache_tensor_bytes": null, "serialized_payload_bytes": 911, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 7938.297540997155, "ttft_ms": 13.029000023379922, "decode_ms": 92.35366608481854, "receiver_prefill_ms": 53.77645790576935}, "total_wall_ms": 8098.523790948093, "ttft_ms": 13.029000023379922, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Koluntek-19", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S1"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "c9648fa6d5cf3d28dcd311f0167cc512f617fded0e15a811bedfd58d5afd86b9", "output_hash": "cefa71c3350f53880f3dfdde8949de2182d8e5275946ac0f441e8068e4a6282a"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "ab422a8b5f9a98a522201209", "attempt_id": "0aa701430be1", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:07:13.036907Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 8}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 422, "message_tokens": 328, "message_utf8_bytes": 911, "cache_tensor_bytes": null, "serialized_payload_bytes": 911, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 7938.297540997155, "ttft_ms": 13.124041957780719, "decode_ms": 13.186166994273663, "receiver_prefill_ms": 52.79804195743054}, "total_wall_ms": 8018.466998939402, "ttft_ms": 13.124041957780719, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S1"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "c9648fa6d5cf3d28dcd311f0167cc512f617fded0e15a811bedfd58d5afd86b9", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a89471fc25cecd9bb6ce4b82", "attempt_id": "c58f30361917", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:07:13.131079Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 8}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 424, "message_tokens": 328, "message_utf8_bytes": 911, "cache_tensor_bytes": null, "serialized_payload_bytes": 911, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 7938.297540997155, "ttft_ms": 13.22887500282377, "decode_ms": 13.134540989995003, "receiver_prefill_ms": 53.54850005824119}, "total_wall_ms": 8019.218957051635, "ttft_ms": 13.22887500282377, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S1"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "c9648fa6d5cf3d28dcd311f0167cc512f617fded0e15a811bedfd58d5afd86b9", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "784da7cf56fe67f9e95c42d8", "attempt_id": "f0b55dec4e8f", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:07:13.225863Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 8}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 424, "message_tokens": 328, "message_utf8_bytes": 911, "cache_tensor_bytes": null, "serialized_payload_bytes": 911, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 7938.297540997155, "ttft_ms": 13.061458012089133, "decode_ms": 13.33450002130121, "receiver_prefill_ms": 53.59429202508181}, "total_wall_ms": 8019.317708094604, "ttft_ms": 13.061458012089133, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S1"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "c9648fa6d5cf3d28dcd311f0167cc512f617fded0e15a811bedfd58d5afd86b9", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "da362a5e87319cf4f812d9d2", "attempt_id": "d77900b72657", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:07:21.239018Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 1138, "message_tokens": 1048, "message_utf8_bytes": 2749, "cache_tensor_bytes": null, "serialized_payload_bytes": 2749, "side_information_bytes": 0, "fetched_bytes": 1688, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 7917.81104193069, "ttft_ms": 13.322749990038574, "decode_ms": 92.72054105531424, "receiver_prefill_ms": 121.62145902402699}, "total_wall_ms": 8147.774875978939, "ttft_ms": 13.322749990038574, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4670383", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S1"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "6432a5195641c31f77de6eb7813b0b720a7ecc19d0ebf6e416f3e8ed09b9aa1e", "output_hash": "19ae8c357e5500bf494a6163bfbd11380b2996e8ea532a2c64f4a599464820f7"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "9a2dec98242f3cd70662a55c", "attempt_id": "09f5ad4151b3", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:07:21.483282Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 1137, "message_tokens": 1044, "message_utf8_bytes": 2696, "cache_tensor_bytes": null, "serialized_payload_bytes": 2696, "side_information_bytes": 0, "fetched_bytes": 1640, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 7917.81104193069, "ttft_ms": 13.3262500166893, "decode_ms": 133.08979105204344, "receiver_prefill_ms": 121.4411249384284}, "total_wall_ms": 8187.446374911815, "ttft_ms": 13.3262500166893, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "NA-58Q-5913", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S1"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a873cd2c7c9f09417957fd91cfdcd0cb29ef5302f1886495f1851f082a01b283", "output_hash": "a4952c0aa4af7f7a21616de82a1eb2bd97ee62e7fa53ca310442860a9bbc6d44"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "9134fdef39601c4cf99473ed", "attempt_id": "9a7e2551e0cb", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:07:21.767027Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 1144, "message_tokens": 1044, "message_utf8_bytes": 2696, "cache_tensor_bytes": null, "serialized_payload_bytes": 2696, "side_information_bytes": 0, "fetched_bytes": 1640, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 7917.81104193069, "ttft_ms": 13.400042080320418, "decode_ms": 13.19354202132672, "receiver_prefill_ms": 121.66991701815277}, "total_wall_ms": 8067.922499845736, "ttft_ms": 13.400042080320418, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S1"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "703bfdb97c7de64338701672f8d43eb4367f3d87efd828858045d70f7bf6e68f", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "cca0e801ba8fe5322fa71fd9", "attempt_id": "5e7c1077fe99", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:07:21.931093Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 1137, "message_tokens": 1044, "message_utf8_bytes": 2696, "cache_tensor_bytes": null, "serialized_payload_bytes": 2696, "side_information_bytes": 0, "fetched_bytes": 1640, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 7917.81104193069, "ttft_ms": 13.514999998733401, "decode_ms": 13.14416597597301, "receiver_prefill_ms": 120.60845794621855}, "total_wall_ms": 8066.887708962895, "ttft_ms": 13.514999998733401, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S1"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "703bfdb97c7de64338701672f8d43eb4367f3d87efd828858045d70f7bf6e68f", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "8915ce3f36a64b1fca280e1e", "attempt_id": "5643847e4ae8", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:07:22.094363Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 1146, "message_tokens": 1044, "message_utf8_bytes": 2696, "cache_tensor_bytes": null, "serialized_payload_bytes": 2696, "side_information_bytes": 0, "fetched_bytes": 1640, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 7917.81104193069, "ttft_ms": 13.391959015280008, "decode_ms": 93.28762500081211, "receiver_prefill_ms": 122.00037494767457}, "total_wall_ms": 8148.338666884229, "ttft_ms": 13.391959015280008, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Koluntek-19", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S1"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "03123c9624d32e03ffa1053a8932d851870f13de8863603788400bb3fe13c630", "output_hash": "cefa71c3350f53880f3dfdde8949de2182d8e5275946ac0f441e8068e4a6282a"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "decf7a325b4b0d46b842afbd", "attempt_id": "67e83b2554ea", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:07:22.338706Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 1137, "message_tokens": 1044, "message_utf8_bytes": 2696, "cache_tensor_bytes": null, "serialized_payload_bytes": 2696, "side_information_bytes": 0, "fetched_bytes": 1640, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 7917.81104193069, "ttft_ms": 13.51412502117455, "decode_ms": 107.51445800997317, "receiver_prefill_ms": 121.56229198444635}, "total_wall_ms": 8162.350666942075, "ttft_ms": 13.51412502117455, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "10444096", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S1"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a83ea03f4edbba54ed269c403db9d64625d20f7ac403618e0255e161228ce8be", "output_hash": "21f7a9931496a7a79f6351211850c8b25960b985ad0ce7449743380582c7c85c"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "2557b796314b31cf88e6e340", "attempt_id": "09485620cb7f", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:07:22.597358Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 1139, "message_tokens": 1044, "message_utf8_bytes": 2696, "cache_tensor_bytes": null, "serialized_payload_bytes": 2696, "side_information_bytes": 0, "fetched_bytes": 1640, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 7917.81104193069, "ttft_ms": 13.441790943033993, "decode_ms": 13.287125038914382, "receiver_prefill_ms": 121.47445802111179}, "total_wall_ms": 8067.861750954762, "ttft_ms": 13.441790943033993, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S1"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a83ea03f4edbba54ed269c403db9d64625d20f7ac403618e0255e161228ce8be", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "d6258df369c9d2abdc170f5f", "attempt_id": "c749652315fb", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:07:22.761450Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 1139, "message_tokens": 1044, "message_utf8_bytes": 2696, "cache_tensor_bytes": null, "serialized_payload_bytes": 2696, "side_information_bytes": 0, "fetched_bytes": 1640, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 7917.81104193069, "ttft_ms": 13.920750003308058, "decode_ms": 13.385499943979084, "receiver_prefill_ms": 129.6718749217689}, "total_wall_ms": 8076.649083872326, "ttft_ms": 13.920750003308058, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S1"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a83ea03f4edbba54ed269c403db9d64625d20f7ac403618e0255e161228ce8be", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c7556bf8bf8713742ec1edc1", "attempt_id": "a79d73b22495", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:07:23.239177Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 87035904}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 44, "message_tokens": null, "message_utf8_bytes": 8303, "cache_tensor_bytes": 87035904, "serialized_payload_bytes": 87035904, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 304.21270802617073, "ttft_ms": 34.21729104593396, "decode_ms": 97.4023750750348, "receiver_prefill_ms": 0.0}, "total_wall_ms": 440.15666597988456, "ttft_ms": 34.21729104593396, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5247196", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S1"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "3ac48b1390a5593419a76c6409f1c54c01affa22edaa1be1850800277310a17c"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a4730701832b93837071abba", "attempt_id": "8327779b037c", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:07:23.389417Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 87035904}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 8303, "cache_tensor_bytes": 87035904, "serialized_payload_bytes": 87035904, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 304.21270802617073, "ttft_ms": 29.8899580957368, "decode_ms": 138.42304202262312, "receiver_prefill_ms": 0.0}, "total_wall_ms": 476.3995830435306, "ttft_ms": 29.8899580957368, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "NA-58Q-5913", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S1"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "a4952c0aa4af7f7a21616de82a1eb2bd97ee62e7fa53ca310442860a9bbc6d44"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "7e3f59410e3be1244fece6a8", "attempt_id": "55edee6aba40", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:07:23.575524Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 87035904}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 54, "message_tokens": null, "message_utf8_bytes": 8303, "cache_tensor_bytes": 87035904, "serialized_payload_bytes": 87035904, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 304.21270802617073, "ttft_ms": 28.68741692509502, "decode_ms": 111.35199991986156, "receiver_prefill_ms": 0.0}, "total_wall_ms": 448.1271239928901, "ttft_ms": 28.68741692509502, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "10047105", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S1"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "bea99610d1e3e75a7f20939faf8e6824e02237b156dacd6442ba2a7e02617510"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a20608235634cdbf760d71c0", "attempt_id": "d70c5862913f", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:07:23.733864Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 87035904}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 8303, "cache_tensor_bytes": 87035904, "serialized_payload_bytes": 87035904, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 304.21270802617073, "ttft_ms": 27.418832993134856, "decode_ms": 96.98529192246497, "receiver_prefill_ms": 0.0}, "total_wall_ms": 431.98395799845457, "ttft_ms": 27.418832993134856, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8227110", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S1"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "85d83b4a8c9cab866c8a91c79511279096225f73c7a6e4226d40a8b0a70020c7"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "f8e52752da1dead80cc47683", "attempt_id": "8aaa4fc0420e", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:07:23.875774Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 87035904}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 56, "message_tokens": null, "message_utf8_bytes": 8303, "cache_tensor_bytes": 87035904, "serialized_payload_bytes": 87035904, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 304.21270802617073, "ttft_ms": 28.59200001694262, "decode_ms": 96.71675006393343, "receiver_prefill_ms": 0.0}, "total_wall_ms": 433.5265001282096, "ttft_ms": 28.59200001694262, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Koluntek-19", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S1"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "cefa71c3350f53880f3dfdde8949de2182d8e5275946ac0f441e8068e4a6282a"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "4817a8e638e1ee33d53c23fc", "attempt_id": "cb20c24bfafb", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:07:24.019441Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 87035904}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 8303, "cache_tensor_bytes": 87035904, "serialized_payload_bytes": 87035904, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 304.21270802617073, "ttft_ms": 27.289582998491824, "decode_ms": 110.60504103079438, "receiver_prefill_ms": 0.0}, "total_wall_ms": 445.8732909988612, "ttft_ms": 27.289582998491824, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "16741.57", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S1"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "894bc3f6d5d2892ccf41fd225103e61b5fff5eeba06784af32b6a52130a222c9"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "ec2d72c91504f925c387519d", "attempt_id": "5bd1eecb73ab", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:07:24.174937Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 87035904}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 49, "message_tokens": null, "message_utf8_bytes": 8303, "cache_tensor_bytes": 87035904, "serialized_payload_bytes": 87035904, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 304.21270802617073, "ttft_ms": 28.055917005985975, "decode_ms": 13.744249939918518, "receiver_prefill_ms": 0.0}, "total_wall_ms": 349.6916249860078, "ttft_ms": 28.055917005985975, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S1"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "9b8c9246dbbddd206cf9adab", "attempt_id": "ad6d1c4daa5f", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:07:24.234722Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 87035904}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 49, "message_tokens": null, "message_utf8_bytes": 8303, "cache_tensor_bytes": 87035904, "serialized_payload_bytes": 87035904, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 304.21270802617073, "ttft_ms": 27.93924999423325, "decode_ms": 13.698333059437573, "receiver_prefill_ms": 0.0}, "total_wall_ms": 349.47424894198775, "ttft_ms": 27.93924999423325, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S1"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "ea82c197b452da7c1ccc8a47", "attempt_id": "bc983d7d95a5", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:07:24.294398Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 2483, "message_tokens": 2397, "message_utf8_bytes": 8833, "cache_tensor_bytes": null, "serialized_payload_bytes": 8833, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.02195802517235279, "ttft_ms": 13.702666037715971, "decode_ms": 120.91924995183945, "receiver_prefill_ms": 270.6754580140114}, "total_wall_ms": 409.1140830423683, "ttft_ms": 13.702666037715971, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4,670,383", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S3"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a5a9d9c66c3e6dd96f2186c32175bb87b92b345b55a43ef3d0fff3c920c7c0bd", "output_hash": "9510633f2e349a1cedecda578083181521646f70ca5ae08bae4b9188f3bd4a3b"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "4387bf8f77729581af8bca62", "attempt_id": "fcca7e359616", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:07:24.717154Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 2486, "message_tokens": 2397, "message_utf8_bytes": 8833, "cache_tensor_bytes": null, "serialized_payload_bytes": 8833, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.02195802517235279, "ttft_ms": 13.529249932616949, "decode_ms": 134.26600000821054, "receiver_prefill_ms": 275.75116709340364}, "total_wall_ms": 427.29949997738004, "ttft_ms": 13.529249932616949, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "NA-58Q-5913", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S3"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a5a9d9c66c3e6dd96f2186c32175bb87b92b345b55a43ef3d0fff3c920c7c0bd", "output_hash": "a4952c0aa4af7f7a21616de82a1eb2bd97ee62e7fa53ca310442860a9bbc6d44"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "dc0f104772c3e03764e72f43", "attempt_id": "d17622163a9b", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:07:25.159027Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 2493, "message_tokens": 2397, "message_utf8_bytes": 8833, "cache_tensor_bytes": null, "serialized_payload_bytes": 8833, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.02195802517235279, "ttft_ms": 13.796708080917597, "decode_ms": 107.77041700202972, "receiver_prefill_ms": 277.4604170117527}, "total_wall_ms": 402.8251660056412, "ttft_ms": 13.796708080917597, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "10047105", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S3"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a5a9d9c66c3e6dd96f2186c32175bb87b92b345b55a43ef3d0fff3c920c7c0bd", "output_hash": "bea99610d1e3e75a7f20939faf8e6824e02237b156dacd6442ba2a7e02617510"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "32b587957d7c46871304c727", "attempt_id": "12cc10296187", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:07:25.576013Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 2486, "message_tokens": 2397, "message_utf8_bytes": 8833, "cache_tensor_bytes": null, "serialized_payload_bytes": 8833, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.02195802517235279, "ttft_ms": 13.81129096262157, "decode_ms": 123.73545789159834, "receiver_prefill_ms": 276.4842079486698}, "total_wall_ms": 417.8327500121668, "ttft_ms": 13.81129096262157, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8,227,110", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S3"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a5a9d9c66c3e6dd96f2186c32175bb87b92b345b55a43ef3d0fff3c920c7c0bd", "output_hash": "c3aa599fbb00899728f00a77f6ec2d981fcc6387f12136cbb65878e6772f4889"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "965f95ea3a4b06c97473bc7f", "attempt_id": "ce440f559131", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:07:26.007791Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 2495, "message_tokens": 2397, "message_utf8_bytes": 8833, "cache_tensor_bytes": null, "serialized_payload_bytes": 8833, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.02195802517235279, "ttft_ms": 13.619250035844743, "decode_ms": 95.29591700993478, "receiver_prefill_ms": 282.9590840265155}, "total_wall_ms": 395.8412909414619, "ttft_ms": 13.619250035844743, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Eshgamgam-11", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S3"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a5a9d9c66c3e6dd96f2186c32175bb87b92b345b55a43ef3d0fff3c920c7c0bd", "output_hash": "8c1313c53ab4dbffa5f336d721c810678133eecaa467241e53d8a92150601475"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "d88a8391bd156c9d416272f8", "attempt_id": "ada91d30ad59", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:07:26.418495Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 2486, "message_tokens": 2397, "message_utf8_bytes": 8833, "cache_tensor_bytes": null, "serialized_payload_bytes": 8833, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.02195802517235279, "ttft_ms": 13.624540995806456, "decode_ms": 136.16520795039833, "receiver_prefill_ms": 287.3835409991443}, "total_wall_ms": 441.1224579671398, "ttft_ms": 13.624540995806456, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "21,467,557", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S3"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a5a9d9c66c3e6dd96f2186c32175bb87b92b345b55a43ef3d0fff3c920c7c0bd", "output_hash": "708dfd38ff9aa40037ed1cb3947eeb539967d7ec47e1982a6a6152905068ce1f"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "28568eb50cdcd4259f05b8f4", "attempt_id": "5bcb3912f130", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:07:26.874208Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 2488, "message_tokens": 2397, "message_utf8_bytes": 8833, "cache_tensor_bytes": null, "serialized_payload_bytes": 8833, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.02195802517235279, "ttft_ms": 13.931749970652163, "decode_ms": 13.68066598661244, "receiver_prefill_ms": 283.33391598425806}, "total_wall_ms": 314.6664999658242, "ttft_ms": 13.931749970652163, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S3"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a5a9d9c66c3e6dd96f2186c32175bb87b92b345b55a43ef3d0fff3c920c7c0bd", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "15ad9ae35ac476f55cb3b624", "attempt_id": "a4dde5bec140", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:07:27.203489Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 2488, "message_tokens": 2397, "message_utf8_bytes": 8833, "cache_tensor_bytes": null, "serialized_payload_bytes": 8833, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.02195802517235279, "ttft_ms": 13.662082958035171, "decode_ms": 13.47633299883455, "receiver_prefill_ms": 298.96795796230435}, "total_wall_ms": 329.8333330312744, "ttft_ms": 13.662082958035171, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S3"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a5a9d9c66c3e6dd96f2186c32175bb87b92b345b55a43ef3d0fff3c920c7c0bd", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "49289f07908f27684631adfd", "attempt_id": "0c6599c58675", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:07:52.763633Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1866}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 1978, "message_tokens": 1891, "message_utf8_bytes": 6446, "cache_tensor_bytes": null, "serialized_payload_bytes": 6446, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25216.283582965843, "ttft_ms": 13.565165922045708, "decode_ms": 119.63337496854365, "receiver_prefill_ms": 207.6283749192953}, "total_wall_ms": 25565.56391599588, "ttft_ms": 13.565165922045708, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4,670,383", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S3"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "0fb8a0eea8884a6977554df160815eb108df87f0a2c2d7623fbd176146fd15c7", "output_hash": "9510633f2e349a1cedecda578083181521646f70ca5ae08bae4b9188f3bd4a3b"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "d30a08d701494252318cfc5a", "attempt_id": "a56bfcebf0b5", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:07:53.127838Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1866}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 1981, "message_tokens": 1891, "message_utf8_bytes": 6446, "cache_tensor_bytes": null, "serialized_payload_bytes": 6446, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25216.283582965843, "ttft_ms": 13.268834096379578, "decode_ms": 133.18262493703514, "receiver_prefill_ms": 206.04487496893853}, "total_wall_ms": 25571.76520791836, "ttft_ms": 13.268834096379578, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "NA-58Q-5913", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S3"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "0fb8a0eea8884a6977554df160815eb108df87f0a2c2d7623fbd176146fd15c7", "output_hash": "a4952c0aa4af7f7a21616de82a1eb2bd97ee62e7fa53ca310442860a9bbc6d44"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "324602294fd4c85df5b8b4fc", "attempt_id": "4da846163072", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:07:53.497601Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1866}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 1988, "message_tokens": 1891, "message_utf8_bytes": 6446, "cache_tensor_bytes": null, "serialized_payload_bytes": 6446, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25216.283582965843, "ttft_ms": 13.764375005848706, "decode_ms": 107.44425002485514, "receiver_prefill_ms": 208.50820804480463}, "total_wall_ms": 25549.129666062072, "ttft_ms": 13.764375005848706, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "10047105", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S3"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "0fb8a0eea8884a6977554df160815eb108df87f0a2c2d7623fbd176146fd15c7", "output_hash": "bea99610d1e3e75a7f20939faf8e6824e02237b156dacd6442ba2a7e02617510"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b2cb1542a3bd6b2a8f1e131f", "attempt_id": "6d9ce6a519f1", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:07:53.844829Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1866}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 1981, "message_tokens": 1891, "message_utf8_bytes": 6446, "cache_tensor_bytes": null, "serialized_payload_bytes": 6446, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25216.283582965843, "ttft_ms": 13.765500043518841, "decode_ms": 94.39554205164313, "receiver_prefill_ms": 232.79533302411437}, "total_wall_ms": 25560.3579999879, "ttft_ms": 13.765500043518841, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8227110", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S3"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "0fb8a0eea8884a6977554df160815eb108df87f0a2c2d7623fbd176146fd15c7", "output_hash": "85d83b4a8c9cab866c8a91c79511279096225f73c7a6e4226d40a8b0a70020c7"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a1c04c3718e4452f5203df7b", "attempt_id": "268f929280d0", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:07:54.203260Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1866}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 1990, "message_tokens": 1891, "message_utf8_bytes": 6446, "cache_tensor_bytes": null, "serialized_payload_bytes": 6446, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25216.283582965843, "ttft_ms": 13.708292040973902, "decode_ms": 94.33608397375792, "receiver_prefill_ms": 233.96616603713483}, "total_wall_ms": 25561.43245799467, "ttft_ms": 13.708292040973902, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Eshgamgam-11", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S3"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "0fb8a0eea8884a6977554df160815eb108df87f0a2c2d7623fbd176146fd15c7", "output_hash": "8c1313c53ab4dbffa5f336d721c810678133eecaa467241e53d8a92150601475"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "692af6aa4d421005f21f13a8", "attempt_id": "fa3de3043d4a", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:07:54.562645Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1866}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 1981, "message_tokens": 1891, "message_utf8_bytes": 6446, "cache_tensor_bytes": null, "serialized_payload_bytes": 6446, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25216.283582965843, "ttft_ms": 13.873500050976872, "decode_ms": 108.18333295173943, "receiver_prefill_ms": 227.0766249857843}, "total_wall_ms": 25568.43329197727, "ttft_ms": 13.873500050976872, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "10965.69", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S3"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "0fb8a0eea8884a6977554df160815eb108df87f0a2c2d7623fbd176146fd15c7", "output_hash": "74adc8ab61809c6e4914bde987b1c6c02e10fd4985b130c0b7bd4523fe12e7fb"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "900d619d464569aef7702542", "attempt_id": "1ab64f1e427f", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:07:54.928732Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1866}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 1983, "message_tokens": 1891, "message_utf8_bytes": 6446, "cache_tensor_bytes": null, "serialized_payload_bytes": 6446, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25216.283582965843, "ttft_ms": 13.668375089764595, "decode_ms": 13.651249930262566, "receiver_prefill_ms": 223.09441701509058}, "total_wall_ms": 25469.625165918842, "ttft_ms": 13.668375089764595, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S3"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "0fb8a0eea8884a6977554df160815eb108df87f0a2c2d7623fbd176146fd15c7", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "497defeb992732b05d99a324", "attempt_id": "d5a5e3add290", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:07:55.196000Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1866}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 1983, "message_tokens": 1891, "message_utf8_bytes": 6446, "cache_tensor_bytes": null, "serialized_payload_bytes": 6446, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25216.283582965843, "ttft_ms": 13.689500046893954, "decode_ms": 13.556374935433269, "receiver_prefill_ms": 234.94866699911654}, "total_wall_ms": 25481.463750009425, "ttft_ms": 13.689500046893954, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S3"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "0fb8a0eea8884a6977554df160815eb108df87f0a2c2d7623fbd176146fd15c7", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "df9f4a98a0071a8f0ae491c6", "attempt_id": "953bd67fdbd8", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:08:23.094476Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 29}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 1236, "message_tokens": 1145, "message_utf8_bytes": 3026, "cache_tensor_bytes": null, "serialized_payload_bytes": 3026, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27619.077584007755, "ttft_ms": 13.399041956290603, "decode_ms": 92.43787499144673, "receiver_prefill_ms": 128.7787080509588}, "total_wall_ms": 27855.54354195483, "ttft_ms": 13.399041956290603, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4670383", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S3"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "22a5303e9183220d9dd2a2160cdd4bb5b8dbb8bd15a523c36c6a1c4c2ed90d04", "output_hash": "19ae8c357e5500bf494a6163bfbd11380b2996e8ea532a2c64f4a599464820f7"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "f25969b32b5d2191f1ec11a1", "attempt_id": "8c980e1eca3d", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:08:23.345717Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 29}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 1239, "message_tokens": 1145, "message_utf8_bytes": 3026, "cache_tensor_bytes": null, "serialized_payload_bytes": 3026, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27619.077584007755, "ttft_ms": 13.300542021170259, "decode_ms": 133.32279189489782, "receiver_prefill_ms": 128.93654103390872}, "total_wall_ms": 27896.46216703113, "ttft_ms": 13.300542021170259, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "NA-58Q-5913", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S3"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "22a5303e9183220d9dd2a2160cdd4bb5b8dbb8bd15a523c36c6a1c4c2ed90d04", "output_hash": "a4952c0aa4af7f7a21616de82a1eb2bd97ee62e7fa53ca310442860a9bbc6d44"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "7470b182cc5a928ce7b83203", "attempt_id": "a01e43f2f8e3", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:08:23.638064Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 29}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 1246, "message_tokens": 1145, "message_utf8_bytes": 3026, "cache_tensor_bytes": null, "serialized_payload_bytes": 3026, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27619.077584007755, "ttft_ms": 13.445832999423146, "decode_ms": 106.3574580475688, "receiver_prefill_ms": 128.75120795797557}, "total_wall_ms": 27869.60479291156, "ttft_ms": 13.445832999423146, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "10047105", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S3"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "22a5303e9183220d9dd2a2160cdd4bb5b8dbb8bd15a523c36c6a1c4c2ed90d04", "output_hash": "bea99610d1e3e75a7f20939faf8e6824e02237b156dacd6442ba2a7e02617510"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "68ecc71da8a4fe9198ebe937", "attempt_id": "7f112d3cfd20", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:08:23.902898Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 29}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 1239, "message_tokens": 1145, "message_utf8_bytes": 3026, "cache_tensor_bytes": null, "serialized_payload_bytes": 3026, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27619.077584007755, "ttft_ms": 13.337124953977764, "decode_ms": 106.44208395387977, "receiver_prefill_ms": 127.97149992547929}, "total_wall_ms": 27868.770709028468, "ttft_ms": 13.337124953977764, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "10047105", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S3"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "22a5303e9183220d9dd2a2160cdd4bb5b8dbb8bd15a523c36c6a1c4c2ed90d04", "output_hash": "bea99610d1e3e75a7f20939faf8e6824e02237b156dacd6442ba2a7e02617510"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "97fd8e4b0f35704dfe323d4e", "attempt_id": "3cce1a5ad27a", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:08:24.166999Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 29}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 1248, "message_tokens": 1145, "message_utf8_bytes": 3026, "cache_tensor_bytes": null, "serialized_payload_bytes": 3026, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27619.077584007755, "ttft_ms": 13.44404206611216, "decode_ms": 93.40012492612004, "receiver_prefill_ms": 127.88579100742936}, "total_wall_ms": 27855.587125988677, "ttft_ms": 13.44404206611216, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Koluntek-19", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S3"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "22a5303e9183220d9dd2a2160cdd4bb5b8dbb8bd15a523c36c6a1c4c2ed90d04", "output_hash": "cefa71c3350f53880f3dfdde8949de2182d8e5275946ac0f441e8068e4a6282a"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "8a1979bf3d68b9b98160879c", "attempt_id": "e363180afb79", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:08:24.418163Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 29}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 1239, "message_tokens": 1145, "message_utf8_bytes": 3026, "cache_tensor_bytes": null, "serialized_payload_bytes": 3026, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27619.077584007755, "ttft_ms": 13.351208064705133, "decode_ms": 93.68995891418308, "receiver_prefill_ms": 128.7109999684617}, "total_wall_ms": 27856.68187495321, "ttft_ms": 13.351208064705133, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1418075", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S3"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "22a5303e9183220d9dd2a2160cdd4bb5b8dbb8bd15a523c36c6a1c4c2ed90d04", "output_hash": "fdb87688b9616571bef4afb8e5a4ac04ff5dfeb0fdca47c9a083751af612c8d6"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b10e65555991592395062b72", "attempt_id": "f9f4c4d5e2ff", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:08:24.670188Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 29}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 1241, "message_tokens": 1145, "message_utf8_bytes": 3026, "cache_tensor_bytes": null, "serialized_payload_bytes": 3026, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27619.077584007755, "ttft_ms": 13.410209096036851, "decode_ms": 13.243917026557028, "receiver_prefill_ms": 128.5649590427056}, "total_wall_ms": 27776.13695897162, "ttft_ms": 13.410209096036851, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S3"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "22a5303e9183220d9dd2a2160cdd4bb5b8dbb8bd15a523c36c6a1c4c2ed90d04", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "3c24ee35662e3e7a7bff0fac", "attempt_id": "5bc9fbadcb44", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:08:24.841674Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 29}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 1241, "message_tokens": 1145, "message_utf8_bytes": 3026, "cache_tensor_bytes": null, "serialized_payload_bytes": 3026, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27619.077584007755, "ttft_ms": 13.563750078901649, "decode_ms": 13.60141602344811, "receiver_prefill_ms": 134.6463750815019}, "total_wall_ms": 27782.678833929822, "ttft_ms": 13.563750078901649, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S3"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "22a5303e9183220d9dd2a2160cdd4bb5b8dbb8bd15a523c36c6a1c4c2ed90d04", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "1f2fc9850c5801d3201d26a3", "attempt_id": "e9c9a6787731", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:08:52.629931Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 2688, "message_tokens": 2598, "message_utf8_bytes": 6678, "cache_tensor_bytes": null, "serialized_payload_bytes": 6678, "side_information_bytes": 0, "fetched_bytes": 3464, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27610.878332983702, "ttft_ms": 13.592417002655566, "decode_ms": 93.89583300799131, "receiver_prefill_ms": 290.8000419847667}, "total_wall_ms": 28015.579998958856, "ttft_ms": 13.592417002655566, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4670383", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S3"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "60a7363eb926959bba8b23530a9657f48daac570828235abfee71a0be3bc80ce", "output_hash": "19ae8c357e5500bf494a6163bfbd11380b2996e8ea532a2c64f4a599464820f7"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "cf2cf075e7f0d1455d47dd5f", "attempt_id": "2fc872e89164", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:08:53.055633Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 2544, "message_tokens": 2451, "message_utf8_bytes": 6392, "cache_tensor_bytes": null, "serialized_payload_bytes": 6392, "side_information_bytes": 0, "fetched_bytes": 3184, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27610.878332983702, "ttft_ms": 13.5754169896245, "decode_ms": 134.01704200077802, "receiver_prefill_ms": 272.08895795047283}, "total_wall_ms": 28034.344290965237, "ttft_ms": 13.5754169896245, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "NA-58Q-5913", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S3"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "6056bba530dd6ca61670606031437048ea73ad366c5d28299508c77b2860a013", "output_hash": "a4952c0aa4af7f7a21616de82a1eb2bd97ee62e7fa53ca310442860a9bbc6d44"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "e16c068695d80d71653a202f", "attempt_id": "8ee208e8c737", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:08:53.494047Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 2693, "message_tokens": 2593, "message_utf8_bytes": 6634, "cache_tensor_bytes": null, "serialized_payload_bytes": 6634, "side_information_bytes": 0, "fetched_bytes": 3439, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27610.878332983702, "ttft_ms": 13.816083082929254, "decode_ms": 108.15900005400181, "receiver_prefill_ms": 309.67616697307676}, "total_wall_ms": 28046.063582994975, "ttft_ms": 13.816083082929254, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "10047105", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S3"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "13cb9c89269866456fd45f9f2126c6609c714cfb8624b65126aa9a63b26c611a", "output_hash": "bea99610d1e3e75a7f20939faf8e6824e02237b156dacd6442ba2a7e02617510"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b3a9148aeebc349f71caaf10", "attempt_id": "9619280f7229", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:08:53.943622Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 2686, "message_tokens": 2593, "message_utf8_bytes": 6634, "cache_tensor_bytes": null, "serialized_payload_bytes": 6634, "side_information_bytes": 0, "fetched_bytes": 3439, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27610.878332983702, "ttft_ms": 14.094625017605722, "decode_ms": 95.96954204607755, "receiver_prefill_ms": 349.3473749840632}, "total_wall_ms": 28073.906790930778, "ttft_ms": 14.094625017605722, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8227110", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S3"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "13cb9c89269866456fd45f9f2126c6609c714cfb8624b65126aa9a63b26c611a", "output_hash": "85d83b4a8c9cab866c8a91c79511279096225f73c7a6e4226d40a8b0a70020c7"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "3900f04f32ceb15debf0a77e", "attempt_id": "ab97b5c981e6", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:08:54.421459Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 2630, "message_tokens": 2528, "message_utf8_bytes": 6513, "cache_tensor_bytes": null, "serialized_payload_bytes": 6513, "side_information_bytes": 0, "fetched_bytes": 3299, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27610.878332983702, "ttft_ms": 76.88708405476063, "decode_ms": 104.23399996943772, "receiver_prefill_ms": 1354.1676660533994}, "total_wall_ms": 29150.06433299277, "ttft_ms": 76.88708405476063, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Koluntek-19", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S3"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "0a557ad40ab1a08012e6da3614678e80355c258cd71b0fac22cd65af2d9b0099", "output_hash": "cefa71c3350f53880f3dfdde8949de2182d8e5275946ac0f441e8068e4a6282a"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "497fb2d2597e572146c6aab5", "attempt_id": "51ce86c871c8", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:08:55.975454Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 2696, "message_tokens": 2603, "message_utf8_bytes": 6662, "cache_tensor_bytes": null, "serialized_payload_bytes": 6662, "side_information_bytes": 0, "fetched_bytes": 3460, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27610.878332983702, "ttft_ms": 13.592334114946425, "decode_ms": 94.30070908274502, "receiver_prefill_ms": 336.325958953239}, "total_wall_ms": 28059.295457904227, "ttft_ms": 13.592334114946425, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1102140", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S3"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "8161aca897e4d9ada52f10c93a924d139c953489f7888d31b2aceaf6dfac7776", "output_hash": "503ad37e945b1bbfcc921d744859e127b6f6cec461c4a83842850c2089c1050a"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "024c3744faac18beb4521255", "attempt_id": "95c09c569bf4", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:08:56.539141Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 2619, "message_tokens": 2524, "message_utf8_bytes": 6603, "cache_tensor_bytes": null, "serialized_payload_bytes": 6603, "side_information_bytes": 0, "fetched_bytes": 3372, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27610.878332983702, "ttft_ms": 14.229790889658034, "decode_ms": 13.849250040948391, "receiver_prefill_ms": 291.79950000252575}, "total_wall_ms": 27934.288374963216, "ttft_ms": 14.229790889658034, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S3"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "da33aa4e5daf7413d35c4063f50525d7536cd3ca1cea7c178702d764cefa00dc", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "12aa12f6a8abee9d13810d09", "attempt_id": "9a4a8ca3bd74", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:08:56.878297Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 2680, "message_tokens": 2585, "message_utf8_bytes": 6621, "cache_tensor_bytes": null, "serialized_payload_bytes": 6621, "side_information_bytes": 0, "fetched_bytes": 3420, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27610.878332983702, "ttft_ms": 13.377916999161243, "decode_ms": 13.602916966192424, "receiver_prefill_ms": 299.9530410161242}, "total_wall_ms": 27941.161875030957, "ttft_ms": 13.377916999161243, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S3"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "bd6415aa332e9f3da313b9a2f21b2d01212046b1ad4bf04a2ff88d2b277ef7b3", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c25abf1109c424f267e340e6", "attempt_id": "3efb45af7fd2", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:08:57.615217Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 94113792}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 44, "message_tokens": null, "message_utf8_bytes": 8779, "cache_tensor_bytes": 94113792, "serialized_payload_bytes": 94113792, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 391.4443749235943, "ttft_ms": 30.72924993466586, "decode_ms": 128.4440839663148, "receiver_prefill_ms": 0.0}, "total_wall_ms": 553.3292499603704, "ttft_ms": 30.72924993466586, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8,497,835", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S3"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "ad5a9b1abadb903c527597b54e67924223db8e556b1863f34086dda7ee84775b"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "db3dd66d0602bb61dff4912c", "attempt_id": "ef1726472165", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:08:57.793134Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 94113792}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 8779, "cache_tensor_bytes": 94113792, "serialized_payload_bytes": 94113792, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 391.4443749235943, "ttft_ms": 28.331374982371926, "decode_ms": 138.63470894284546, "receiver_prefill_ms": 0.0}, "total_wall_ms": 561.5885419538245, "ttft_ms": 28.331374982371926, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "NA-58Q-5913", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S3"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "a4952c0aa4af7f7a21616de82a1eb2bd97ee62e7fa53ca310442860a9bbc6d44"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "323440943fa4c0dcdf5bce78", "attempt_id": "65dbfff3e394", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:08:57.978106Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 94113792}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 54, "message_tokens": null, "message_utf8_bytes": 8779, "cache_tensor_bytes": 94113792, "serialized_payload_bytes": 94113792, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 391.4443749235943, "ttft_ms": 29.894166975282133, "decode_ms": 97.5865840446204, "receiver_prefill_ms": 0.0}, "total_wall_ms": 522.1229169983417, "ttft_ms": 29.894166975282133, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4932140", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S3"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "7e87542fa4bdc4ea71fe8d69f229e6c3d83198d33a51aab51951229278d6dea5"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "495a08cbf2459b0b1780f35c", "attempt_id": "4f45d6bf1aae", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:08:58.123489Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 94113792}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 8779, "cache_tensor_bytes": 94113792, "serialized_payload_bytes": 94113792, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 391.4443749235943, "ttft_ms": 28.16562505904585, "decode_ms": 125.1927079865709, "receiver_prefill_ms": 0.0}, "total_wall_ms": 547.9260408319533, "ttft_ms": 28.16562505904585, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8,227,110", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S3"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "c3aa599fbb00899728f00a77f6ec2d981fcc6387f12136cbb65878e6772f4889"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "6948a460659cd63eb5843a81", "attempt_id": "8a75ccbf64dc", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:08:58.294570Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 94113792}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 56, "message_tokens": null, "message_utf8_bytes": 8779, "cache_tensor_bytes": 94113792, "serialized_payload_bytes": 94113792, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 391.4443749235943, "ttft_ms": 30.134665896184742, "decode_ms": 96.74820804502815, "receiver_prefill_ms": 0.0}, "total_wall_ms": 521.4604159118608, "ttft_ms": 30.134665896184742, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Koluntek-19", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S3"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "cefa71c3350f53880f3dfdde8949de2182d8e5275946ac0f441e8068e4a6282a"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "6df9c76e8a3c70ef8c291086", "attempt_id": "eb76c56de7f2", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:08:58.438854Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 94113792}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 8779, "cache_tensor_bytes": 94113792, "serialized_payload_bytes": 94113792, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 391.4443749235943, "ttft_ms": 28.230625088326633, "decode_ms": 138.109749997966, "receiver_prefill_ms": 0.0}, "total_wall_ms": 560.9433328500018, "ttft_ms": 28.230625088326633, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "11,046,021", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S3"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "33a22657f9b76254a43b7c62101c70023084c4914237582246c9d4ec4d664a14"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "016732fc5dd933240abbc58f", "attempt_id": "cb77ac1671d2", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:08:58.622301Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 94113792}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 49, "message_tokens": null, "message_utf8_bytes": 8779, "cache_tensor_bytes": 94113792, "serialized_payload_bytes": 94113792, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 391.4443749235943, "ttft_ms": 29.547999962233007, "decode_ms": 138.3587090531364, "receiver_prefill_ms": 0.0}, "total_wall_ms": 562.6298749120906, "ttft_ms": 29.547999962233007, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "GY-60C-8075", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S3"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "d2c9b339e98f144ecc49f158aa378ba0adac98b86dfe44933a71225d12ec1e93"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "277bf80296ba18dc6426a408", "attempt_id": "184f4d4a0967", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:08:58.808322Z", "task_family": "synthetic", "document_id": "syn_03009_2000", "document_hash": "6ed94dcf728255bd7a7d11c97c362930728fb81a3e91f1906eb6eec86e0fedb7", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03009_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 94113792}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1976, "query_tokens": 49, "message_tokens": null, "message_utf8_bytes": 8779, "cache_tensor_bytes": 94113792, "serialized_payload_bytes": 94113792, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 391.4443749235943, "ttft_ms": 29.01087503414601, "decode_ms": 13.880333048291504, "receiver_prefill_ms": 0.0}, "total_wall_ms": 437.4123748857528, "ttft_ms": 29.01087503414601, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Felker-16", "bogus_amount": "4,775,305", "source": "S3"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03009_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c00cec166900566486d637e1", "attempt_id": "fef9baec006d", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:08:59.782148Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 6165, "message_tokens": 6079, "message_utf8_bytes": 30047, "cache_tensor_bytes": null, "serialized_payload_bytes": 30047, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.03420806024223566, "ttft_ms": 13.827749993652105, "decode_ms": 95.33779206685722, "receiver_prefill_ms": 732.7668749494478}, "total_wall_ms": 853.3531250432134, "ttft_ms": 13.827749993652105, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "892,005", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "770f27c22ee7dab2f388c262ed57b5770f0979f5e4b6226b2451d6d3e057af82", "output_hash": "11a79dcb724bdfa4933fdc68ce7331d64c8d288cc386b3ccf051af4279b5d355"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "f30bf7a1473399cc4cfd35d5", "attempt_id": "672e7998d625", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:09:00.652410Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 6168, "message_tokens": 6079, "message_utf8_bytes": 30047, "cache_tensor_bytes": null, "serialized_payload_bytes": 30047, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.03420806024223566, "ttft_ms": 13.619665987789631, "decode_ms": 135.79141697846353, "receiver_prefill_ms": 720.5675840377808}, "total_wall_ms": 879.4797910377383, "ttft_ms": 13.619665987789631, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "XX-43W-9874", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "770f27c22ee7dab2f388c262ed57b5770f0979f5e4b6226b2451d6d3e057af82", "output_hash": "0ab6d0ab000bed72bda142906dbe114882d3e55d9e45808c6f2ae838a7201194"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "70f640ee3ed05910ffced14d", "attempt_id": "bca7b7e0c3ed", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:09:01.550801Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 6172, "message_tokens": 6079, "message_utf8_bytes": 30047, "cache_tensor_bytes": null, "serialized_payload_bytes": 30047, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.03420806024223566, "ttft_ms": 14.083584072068334, "decode_ms": 82.61400007177144, "receiver_prefill_ms": 719.9002089910209}, "total_wall_ms": 826.0931660188362, "ttft_ms": 14.083584072068334, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "931295", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "770f27c22ee7dab2f388c262ed57b5770f0979f5e4b6226b2451d6d3e057af82", "output_hash": "0291810ef629db45b8dd44dbd2ffb05f6ec39daa8fe059b1ab9f757164cabc41"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "517ab57fbf1284863527c033", "attempt_id": "cdc91a9db957", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:09:02.394357Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 6165, "message_tokens": 6079, "message_utf8_bytes": 30047, "cache_tensor_bytes": null, "serialized_payload_bytes": 30047, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.03420806024223566, "ttft_ms": 13.683542027138174, "decode_ms": 122.49920808244497, "receiver_prefill_ms": 720.5650829710066}, "total_wall_ms": 866.2508750567213, "ttft_ms": 13.683542027138174, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "3,875,883", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "770f27c22ee7dab2f388c262ed57b5770f0979f5e4b6226b2451d6d3e057af82", "output_hash": "a8243b862e2d82445ca3c93eec5e829e5d41cfa4d9abfba7b18d922fdebba636"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "d247e70c48ce88b2112a10d0", "attempt_id": "300fdfee01e7", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:09:03.279929Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 6175, "message_tokens": 6079, "message_utf8_bytes": 30047, "cache_tensor_bytes": null, "serialized_payload_bytes": 30047, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.03420806024223566, "ttft_ms": 13.850750052370131, "decode_ms": 82.2020840132609, "receiver_prefill_ms": 732.8845410374925}, "total_wall_ms": 838.7763330247253, "ttft_ms": 13.850750052370131, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Rukmelri-6", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "770f27c22ee7dab2f388c262ed57b5770f0979f5e4b6226b2451d6d3e057af82", "output_hash": "2e6780effd3ce64b936cdd7bf5b52ba071dcd639cc48639e8f72d04cc6978a32"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "ad5815c95e9d644786ae0bb9", "attempt_id": "3b5a9b13fef8", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:09:04.135775Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 6168, "message_tokens": 6079, "message_utf8_bytes": 30047, "cache_tensor_bytes": null, "serialized_payload_bytes": 30047, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.03420806024223566, "ttft_ms": 13.96854198537767, "decode_ms": 137.0744169689715, "receiver_prefill_ms": 747.6423340849578}, "total_wall_ms": 908.3669580286369, "ttft_ms": 13.96854198537767, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "23,446,718", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "770f27c22ee7dab2f388c262ed57b5770f0979f5e4b6226b2451d6d3e057af82", "output_hash": "a91538823f6151ced3a46d7efa81097fa118faf780ed073f0919a2aa342abbbe"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a2129cd8da002d4bc8042365", "attempt_id": "ad45fc52531a", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:09:05.061076Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 6168, "message_tokens": 6079, "message_utf8_bytes": 30047, "cache_tensor_bytes": null, "serialized_payload_bytes": 30047, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.03420806024223566, "ttft_ms": 14.023917028680444, "decode_ms": 13.82945803925395, "receiver_prefill_ms": 758.4157079691067}, "total_wall_ms": 795.6905830651522, "ttft_ms": 14.023917028680444, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "770f27c22ee7dab2f388c262ed57b5770f0979f5e4b6226b2451d6d3e057af82", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "f4542da47ba544fa01705937", "attempt_id": "82185415f1ea", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:09:05.873076Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 6170, "message_tokens": 6079, "message_utf8_bytes": 30047, "cache_tensor_bytes": null, "serialized_payload_bytes": 30047, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.03420806024223566, "ttft_ms": 14.009708072990179, "decode_ms": 13.857541023753583, "receiver_prefill_ms": 796.3352500228211}, "total_wall_ms": 833.7395001435652, "ttft_ms": 14.009708072990179, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "770f27c22ee7dab2f388c262ed57b5770f0979f5e4b6226b2451d6d3e057af82", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "582489a5ca335ee0536d0663", "attempt_id": "7122dda4c94c", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:09:21.212326Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 1125, "message_tokens": 1037, "message_utf8_bytes": 2792, "cache_tensor_bytes": null, "serialized_payload_bytes": 2792, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14488.143707974814, "ttft_ms": 13.417957932688296, "decode_ms": 80.583417089656, "receiver_prefill_ms": 122.18870804645121}, "total_wall_ms": 14706.926958053373, "ttft_ms": 13.417957932688296, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "892005", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "0f06f0902a3365d6a081773e6e5d0b3c1dc1d85efd4c2cc62ad9d23ef1b71413", "output_hash": "296832ce0f2ee0fb1dcf407ccce524ec73af796101cbe85b60c358235bf11a7e"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "6bf96b65216379d9eb980325", "attempt_id": "33c54b45073c", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:09:21.448773Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 1128, "message_tokens": 1037, "message_utf8_bytes": 2792, "cache_tensor_bytes": null, "serialized_payload_bytes": 2792, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14488.143707974814, "ttft_ms": 13.297833036631346, "decode_ms": 13.32679099868983, "receiver_prefill_ms": 121.71629199292511}, "total_wall_ms": 14638.394874986261, "ttft_ms": 13.297833036631346, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "0f06f0902a3365d6a081773e6e5d0b3c1dc1d85efd4c2cc62ad9d23ef1b71413", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "6381aa88d1b7c8789aa01a61", "attempt_id": "36b8bb27f0fc", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:09:21.616483Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 1132, "message_tokens": 1037, "message_utf8_bytes": 2792, "cache_tensor_bytes": null, "serialized_payload_bytes": 2792, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14488.143707974814, "ttft_ms": 13.380000018514693, "decode_ms": 79.77225002832711, "receiver_prefill_ms": 122.04491591546685}, "total_wall_ms": 14705.275415908545, "ttft_ms": 13.380000018514693, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "931295", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "0f06f0902a3365d6a081773e6e5d0b3c1dc1d85efd4c2cc62ad9d23ef1b71413", "output_hash": "0291810ef629db45b8dd44dbd2ffb05f6ec39daa8fe059b1ab9f757164cabc41"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "7d8ccc4675a563150a0d5ed3", "attempt_id": "9c2058c3870c", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:09:21.850955Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 1125, "message_tokens": 1037, "message_utf8_bytes": 2792, "cache_tensor_bytes": null, "serialized_payload_bytes": 2792, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14488.143707974814, "ttft_ms": 13.325708918273449, "decode_ms": 93.16733409650624, "receiver_prefill_ms": 121.27762497402728}, "total_wall_ms": 14717.777374899015, "ttft_ms": 13.325708918273449, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "3875800", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "0f06f0902a3365d6a081773e6e5d0b3c1dc1d85efd4c2cc62ad9d23ef1b71413", "output_hash": "559106be0588d73a6b3047f0d477e9db4470b10a22266d679f7a968d97bf5ab6"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "64675ae30f78470d982cfae8", "attempt_id": "81c1fd071157", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:09:22.097453Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 1135, "message_tokens": 1037, "message_utf8_bytes": 2792, "cache_tensor_bytes": null, "serialized_payload_bytes": 2792, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14488.143707974814, "ttft_ms": 13.325749896466732, "decode_ms": 79.54175001941621, "receiver_prefill_ms": 121.47166708018631}, "total_wall_ms": 14704.394040978514, "ttft_ms": 13.325749896466732, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Rukmelri-6", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "0f06f0902a3365d6a081773e6e5d0b3c1dc1d85efd4c2cc62ad9d23ef1b71413", "output_hash": "2e6780effd3ce64b936cdd7bf5b52ba071dcd639cc48639e8f72d04cc6978a32"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a40f6643b4255be240d4f8ed", "attempt_id": "8a9815f08b3a", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:09:22.330691Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 1128, "message_tokens": 1037, "message_utf8_bytes": 2792, "cache_tensor_bytes": null, "serialized_payload_bytes": 2792, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14488.143707974814, "ttft_ms": 13.185792020522058, "decode_ms": 132.32904206961393, "receiver_prefill_ms": 121.2509999750182}, "total_wall_ms": 14756.762541946955, "ttft_ms": 13.185792020522058, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "34,582,865", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "0f06f0902a3365d6a081773e6e5d0b3c1dc1d85efd4c2cc62ad9d23ef1b71413", "output_hash": "16fe870e482498ba710284c408d272dd4c267776c517d14c8a4e78269bda543f"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "272c622954dbbad61e250489", "attempt_id": "3fab03ceca30", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:09:22.616035Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 1128, "message_tokens": 1037, "message_utf8_bytes": 2792, "cache_tensor_bytes": null, "serialized_payload_bytes": 2792, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14488.143707974814, "ttft_ms": 13.28833308070898, "decode_ms": 13.40516610071063, "receiver_prefill_ms": 121.03675003163517}, "total_wall_ms": 14637.831291998737, "ttft_ms": 13.28833308070898, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "0f06f0902a3365d6a081773e6e5d0b3c1dc1d85efd4c2cc62ad9d23ef1b71413", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "d1635deeadde3b527e2d9e52", "attempt_id": "716d9bacca67", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:09:22.782827Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 1130, "message_tokens": 1037, "message_utf8_bytes": 2792, "cache_tensor_bytes": null, "serialized_payload_bytes": 2792, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14488.143707974814, "ttft_ms": 13.428666978143156, "decode_ms": 13.26979196164757, "receiver_prefill_ms": 121.35370797477663}, "total_wall_ms": 14638.103291974403, "ttft_ms": 13.428666978143156, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "0f06f0902a3365d6a081773e6e5d0b3c1dc1d85efd4c2cc62ad9d23ef1b71413", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "76b8eeb22e23f345c28f48c4", "attempt_id": "9da47b6b2dc6", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:09:37.522601Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 14}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 649, "message_tokens": 558, "message_utf8_bytes": 1557, "cache_tensor_bytes": null, "serialized_payload_bytes": 1557, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14572.15816702228, "ttft_ms": 13.125375029630959, "decode_ms": 13.354166992940009, "receiver_prefill_ms": 75.17187495250255}, "total_wall_ms": 14675.064667011611, "ttft_ms": 13.125375029630959, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "e9c6c68daaec68c1c50e85ae2893c8126eb3ec644e0fb30f4d89c937ed86a89f", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "ff2f176811c05a2f60b3c9e1", "attempt_id": "6c23a2011feb", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:09:37.642996Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 14}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 652, "message_tokens": 558, "message_utf8_bytes": 1557, "cache_tensor_bytes": null, "serialized_payload_bytes": 1557, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14572.15816702228, "ttft_ms": 13.321875012479722, "decode_ms": 13.42945802025497, "receiver_prefill_ms": 75.05762507207692}, "total_wall_ms": 14675.306875025854, "ttft_ms": 13.321875012479722, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "e9c6c68daaec68c1c50e85ae2893c8126eb3ec644e0fb30f4d89c937ed86a89f", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c3d33c834c10015de6c92474", "attempt_id": "c0ec4fd8e420", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:09:37.763589Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 14}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 656, "message_tokens": 558, "message_utf8_bytes": 1557, "cache_tensor_bytes": null, "serialized_payload_bytes": 1557, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14572.15816702228, "ttft_ms": 13.311874936334789, "decode_ms": 13.12079094350338, "receiver_prefill_ms": 74.91662504617125}, "total_wall_ms": 14674.838876002468, "ttft_ms": 13.311874936334789, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "e9c6c68daaec68c1c50e85ae2893c8126eb3ec644e0fb30f4d89c937ed86a89f", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "56e9cd06d892bde7d02b5ba3", "attempt_id": "52bdb2d4f2b9", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:09:37.883664Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 14}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 649, "message_tokens": 558, "message_utf8_bytes": 1557, "cache_tensor_bytes": null, "serialized_payload_bytes": 1557, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14572.15816702228, "ttft_ms": 13.596458942629397, "decode_ms": 92.59366698097438, "receiver_prefill_ms": 74.37675003893673}, "total_wall_ms": 14753.9927510079, "ttft_ms": 13.596458942629397, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5735925", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "e9c6c68daaec68c1c50e85ae2893c8126eb3ec644e0fb30f4d89c937ed86a89f", "output_hash": "166e88de295f66164de6a78d1ccebde4c07b308215fcefdfa0acb358c02bef6e"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "460b7fd84e3cb291dc5a3135", "attempt_id": "54be0fa382c4", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:09:38.082779Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 14}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 659, "message_tokens": 558, "message_utf8_bytes": 1557, "cache_tensor_bytes": null, "serialized_payload_bytes": 1557, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14572.15816702228, "ttft_ms": 13.253374956548214, "decode_ms": 79.852957976982, "receiver_prefill_ms": 75.08829096332192}, "total_wall_ms": 14741.642125998624, "ttft_ms": 13.253374956548214, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Rukmelri-6", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "e9c6c68daaec68c1c50e85ae2893c8126eb3ec644e0fb30f4d89c937ed86a89f", "output_hash": "2e6780effd3ce64b936cdd7bf5b52ba071dcd639cc48639e8f72d04cc6978a32"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "88439de402616e1e5f366f06", "attempt_id": "80ed1a6e5f29", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:09:38.269257Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 14}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 652, "message_tokens": 558, "message_utf8_bytes": 1557, "cache_tensor_bytes": null, "serialized_payload_bytes": 1557, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14572.15816702228, "ttft_ms": 13.288083020597696, "decode_ms": 133.11179098673165, "receiver_prefill_ms": 74.59700002800673}, "total_wall_ms": 14794.453542097472, "ttft_ms": 13.288083020597696, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "16,515,620", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "e9c6c68daaec68c1c50e85ae2893c8126eb3ec644e0fb30f4d89c937ed86a89f", "output_hash": "c0a0020e631e18604028d0119ba2d1d6103ae9d9bdcf54c720cc976926f869cb"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b4789b5c896bdb673570a1b4", "attempt_id": "677355813973", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:09:38.511454Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 14}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 652, "message_tokens": 558, "message_utf8_bytes": 1557, "cache_tensor_bytes": null, "serialized_payload_bytes": 1557, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14572.15816702228, "ttft_ms": 13.329625013284385, "decode_ms": 13.010332942940295, "receiver_prefill_ms": 74.24270804040134}, "total_wall_ms": 14674.029999994673, "ttft_ms": 13.329625013284385, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "e9c6c68daaec68c1c50e85ae2893c8126eb3ec644e0fb30f4d89c937ed86a89f", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "81e8a04e43ccadc6e1a660e9", "attempt_id": "21d01374ac8c", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:09:38.630991Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 14}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 654, "message_tokens": 558, "message_utf8_bytes": 1557, "cache_tensor_bytes": null, "serialized_payload_bytes": 1557, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14572.15816702228, "ttft_ms": 13.16612504888326, "decode_ms": 13.120624935254455, "receiver_prefill_ms": 74.97362501453608}, "total_wall_ms": 14674.677959061228, "ttft_ms": 13.16612504888326, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "e9c6c68daaec68c1c50e85ae2893c8126eb3ec644e0fb30f4d89c937ed86a89f", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "90bdd8d603758162b4d5b3b4", "attempt_id": "17b1c5d51765", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:09:53.304223Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 1372, "message_tokens": 1282, "message_utf8_bytes": 3456, "cache_tensor_bytes": null, "serialized_payload_bytes": 3456, "side_information_bytes": 0, "fetched_bytes": 1789, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14549.023207975551, "ttft_ms": 13.31787509843707, "decode_ms": 80.93829196877778, "receiver_prefill_ms": 145.06479201372713}, "total_wall_ms": 14790.541582973674, "ttft_ms": 13.31787509843707, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "892005", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "6e3be9dcb4d4321fa5d15f9f9008de489d11190d7947cac896aad836a28b31c1", "output_hash": "296832ce0f2ee0fb1dcf407ccce524ec73af796101cbe85b60c358235bf11a7e"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a5149e52c42976c388893dc7", "attempt_id": "e8cb997c8622", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:09:53.565785Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 1351, "message_tokens": 1258, "message_utf8_bytes": 3325, "cache_tensor_bytes": null, "serialized_payload_bytes": 3325, "side_information_bytes": 0, "fetched_bytes": 1666, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14549.023207975551, "ttft_ms": 13.549334020353854, "decode_ms": 13.877624995075166, "receiver_prefill_ms": 144.18000006116927}, "total_wall_ms": 14723.153042024933, "ttft_ms": 13.549334020353854, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a68eaed2d8e20be60b6b63eaf9c2915a466015339ecc7aa7863b81737a793882", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "3df7d1b414a60e381f9a3b65", "attempt_id": "1d33e7f9f826", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:09:53.757484Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 1385, "message_tokens": 1288, "message_utf8_bytes": 3417, "cache_tensor_bytes": null, "serialized_payload_bytes": 3417, "side_information_bytes": 0, "fetched_bytes": 1750, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14549.023207975551, "ttft_ms": 13.383416924625635, "decode_ms": 13.768707984127104, "receiver_prefill_ms": 148.4429578995332}, "total_wall_ms": 14726.763082901016, "ttft_ms": 13.383416924625635, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "66c6dd05e24dc88b5474ef964e8540967f0eb165b89aca5b6a097f8d94381d86", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "f283dd25b6675baa9e0c13bc", "attempt_id": "f5e14d380d04", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:09:53.952015Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 1378, "message_tokens": 1288, "message_utf8_bytes": 3417, "cache_tensor_bytes": null, "serialized_payload_bytes": 3417, "side_information_bytes": 0, "fetched_bytes": 1750, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14549.023207975551, "ttft_ms": 13.32008303143084, "decode_ms": 93.34087499883026, "receiver_prefill_ms": 147.66699995379895}, "total_wall_ms": 14805.56154192891, "ttft_ms": 13.32008303143084, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "6167665", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "66c6dd05e24dc88b5474ef964e8540967f0eb165b89aca5b6a097f8d94381d86", "output_hash": "d84a136af40fdfbf5493b7cd9e968e853ee36bbccb7d2b0154a0f417ff569de5"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "3dd730ad3db9e314b9b9ee54", "attempt_id": "c2067c7e5fec", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:09:54.225613Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 1351, "message_tokens": 1251, "message_utf8_bytes": 3334, "cache_tensor_bytes": null, "serialized_payload_bytes": 3334, "side_information_bytes": 0, "fetched_bytes": 1682, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14549.023207975551, "ttft_ms": 13.452165992930532, "decode_ms": 93.06245797779411, "receiver_prefill_ms": 144.5826250128448}, "total_wall_ms": 14802.774415933527, "ttft_ms": 13.452165992930532, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Project Rukmelri-6", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 0.6666666666666666}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a1b6f76bffcd3428dad54268b3e6621f3fe4004b287a45d51bf6ec944c3e3eba", "output_hash": "082f80c07a98bf86169973431a67304ef9ffc0f5700900613d4021b4262f5e8a"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "3d67387f91d63ee515bc12cf", "attempt_id": "1095df0b49b5", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:09:54.496906Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 1375, "message_tokens": 1282, "message_utf8_bytes": 3456, "cache_tensor_bytes": null, "serialized_payload_bytes": 3456, "side_information_bytes": 0, "fetched_bytes": 1789, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14549.023207975551, "ttft_ms": 13.634208007715642, "decode_ms": 135.21937502082437, "receiver_prefill_ms": 146.91200002562255}, "total_wall_ms": 14846.945999888703, "ttft_ms": 13.634208007715642, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "15,475,625", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "5d4e155f1ec6666eef915c6b958db95b1593d00c5f8f98bbdfe239ea8afed18d", "output_hash": "a66a7019b07a681fb79a12db9932c181aac31a482b13ee950d0fd58f77fc2956"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "55eb5edb9578a0c03c01de1f", "attempt_id": "902ca81415a5", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:09:54.811811Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 1370, "message_tokens": 1277, "message_utf8_bytes": 3400, "cache_tensor_bytes": null, "serialized_payload_bytes": 3400, "side_information_bytes": 0, "fetched_bytes": 1731, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14549.023207975551, "ttft_ms": 13.606499996967614, "decode_ms": 13.498499989509583, "receiver_prefill_ms": 149.617874994874}, "total_wall_ms": 14727.864207932726, "ttft_ms": 13.606499996967614, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "37af9f3a473dcd11125daef928bb2cedebc161cf9bf3caeed66ad5a0ea80881d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "bd12a5f0a53c7f56080e77e8", "attempt_id": "fe38d2296a61", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:09:55.007502Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 1383, "message_tokens": 1288, "message_utf8_bytes": 3417, "cache_tensor_bytes": null, "serialized_payload_bytes": 3417, "side_information_bytes": 0, "fetched_bytes": 1750, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14549.023207975551, "ttft_ms": 13.499332941137254, "decode_ms": 13.340250006876886, "receiver_prefill_ms": 159.8149590427056}, "total_wall_ms": 14737.819666974247, "ttft_ms": 13.499332941137254, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "b1128198ab4726c238cc12e5fd86b98ea06befc51f9d744dcf452df8935a1b6d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "fc8af74796f4691adb6cb77d", "attempt_id": "419caa7f40bd", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:09:56.032251Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 226971648}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 44, "message_tokens": null, "message_utf8_bytes": 30021, "cache_tensor_bytes": 226971648, "serialized_payload_bytes": 226971648, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 818.5930420877412, "ttft_ms": 55.827916017733514, "decode_ms": 90.9731670981273, "receiver_prefill_ms": 0.0}, "total_wall_ms": 974.6370840584859, "ttft_ms": 55.827916017733514, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "892005", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "296832ce0f2ee0fb1dcf407ccce524ec73af796101cbe85b60c358235bf11a7e"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "842353a167b2b70c1ea8c8fc", "attempt_id": "708810e5f038", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:09:56.205612Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 226971648}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 30021, "cache_tensor_bytes": 226971648, "serialized_payload_bytes": 226971648, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 818.5930420877412, "ttft_ms": 42.147250031121075, "decode_ms": 150.2298340201378, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1018.5367921367288, "ttft_ms": 42.147250031121075, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "XX-43W-9874", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "0ab6d0ab000bed72bda142906dbe114882d3e55d9e45808c6f2ae838a7201194"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "62845284b3ef9a11fa7b15de", "attempt_id": "3adae2ad8147", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:09:56.422581Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 226971648}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 51, "message_tokens": null, "message_utf8_bytes": 30021, "cache_tensor_bytes": 226971648, "serialized_payload_bytes": 226971648, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 818.5930420877412, "ttft_ms": 42.237707995809615, "decode_ms": 89.92499997839332, "receiver_prefill_ms": 0.0}, "total_wall_ms": 958.2848760765046, "ttft_ms": 42.237707995809615, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "931295", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "0291810ef629db45b8dd44dbd2ffb05f6ec39daa8fe059b1ab9f757164cabc41"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "0dbbf82f2c21123dce07e144", "attempt_id": "aa434d57d6a7", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:09:56.579835Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 226971648}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 44, "message_tokens": null, "message_utf8_bytes": 30021, "cache_tensor_bytes": 226971648, "serialized_payload_bytes": 226971648, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 818.5930420877412, "ttft_ms": 48.64433303009719, "decode_ms": 134.01350006461143, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1008.511167136021, "ttft_ms": 48.64433303009719, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "3,418,930", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "0c81bdc5a7cd40d6d4b0c52dc5fc252e6a9579c7b15fd2d5c7cfa4590ccab621"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "f0e91a1c7b3f6be689411734", "attempt_id": "37ebd019583c", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:09:56.787741Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 226971648}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 54, "message_tokens": null, "message_utf8_bytes": 30021, "cache_tensor_bytes": 226971648, "serialized_payload_bytes": 226971648, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 818.5930420877412, "ttft_ms": 47.444958006963134, "decode_ms": 89.27520900033414, "receiver_prefill_ms": 0.0}, "total_wall_ms": 962.8057511290535, "ttft_ms": 47.444958006963134, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Rukmelri-6", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "2e6780effd3ce64b936cdd7bf5b52ba071dcd639cc48639e8f72d04cc6978a32"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "61e8bedc9354400daab3732b", "attempt_id": "1a80484b0eef", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:09:56.949394Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 226971648}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 30021, "cache_tensor_bytes": 226971648, "serialized_payload_bytes": 226971648, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 818.5930420877412, "ttft_ms": 41.96329100523144, "decode_ms": 149.18712503276765, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1017.1880000270903, "ttft_ms": 41.96329100523144, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "25,185,613", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "da5883ecb84edec81fa606626e1729a23a58879bb7de429ba6f1c24467bacc56"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "e76097a095f901c9480d7afd", "attempt_id": "6d76e56c549d", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:09:57.165726Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 226971648}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 30021, "cache_tensor_bytes": 226971648, "serialized_payload_bytes": 226971648, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 818.5930420877412, "ttft_ms": 41.95737501140684, "decode_ms": 14.907749951817095, "receiver_prefill_ms": 0.0}, "total_wall_ms": 882.9420839902014, "ttft_ms": 41.95737501140684, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "bcb5f81e113509fc6f0b40fe", "attempt_id": "8ad0d5350e7a", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:09:57.247186Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 226971648}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 49, "message_tokens": null, "message_utf8_bytes": 30021, "cache_tensor_bytes": 226971648, "serialized_payload_bytes": 226971648, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 818.5930420877412, "ttft_ms": 45.07395892869681, "decode_ms": 15.00237500295043, "receiver_prefill_ms": 0.0}, "total_wall_ms": 886.1217501107603, "ttft_ms": 45.07395892869681, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "82c69fe75185afb7cf662b18", "attempt_id": "2e178953f016", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:09:57.331321Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 6279, "message_tokens": 6193, "message_utf8_bytes": 30551, "cache_tensor_bytes": null, "serialized_payload_bytes": 30551, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.026957946829497814, "ttft_ms": 14.125708024948835, "decode_ms": 95.87779105640948, "receiver_prefill_ms": 762.0474169962108}, "total_wall_ms": 886.327374028042, "ttft_ms": 14.125708024948835, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "892,005", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "bc7084590d7e4773b7e90ae4a9c7b9ead88f46214128f9ec5b71b9487bf56e9e", "output_hash": "11a79dcb724bdfa4933fdc68ce7331d64c8d288cc386b3ccf051af4279b5d355"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "45d1c7920275d69becbcbf20", "attempt_id": "ee4a0599f253", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:09:58.234056Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 6282, "message_tokens": 6193, "message_utf8_bytes": 30551, "cache_tensor_bytes": null, "serialized_payload_bytes": 30551, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.026957946829497814, "ttft_ms": 14.060958987101912, "decode_ms": 137.95495906379074, "receiver_prefill_ms": 780.9259579516947}, "total_wall_ms": 942.891790997237, "ttft_ms": 14.060958987101912, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "XX-43W-9874", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "bc7084590d7e4773b7e90ae4a9c7b9ead88f46214128f9ec5b71b9487bf56e9e", "output_hash": "0ab6d0ab000bed72bda142906dbe114882d3e55d9e45808c6f2ae838a7201194"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "26f7b2664958356bd6c5e4a1", "attempt_id": "ffe0fd17b3a9", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:09:59.194222Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 6286, "message_tokens": 6193, "message_utf8_bytes": 30551, "cache_tensor_bytes": null, "serialized_payload_bytes": 30551, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.026957946829497814, "ttft_ms": 15.072999987751245, "decode_ms": 84.04408395290375, "receiver_prefill_ms": 795.2345410594717}, "total_wall_ms": 904.021373949945, "ttft_ms": 15.072999987751245, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "931295", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "bc7084590d7e4773b7e90ae4a9c7b9ead88f46214128f9ec5b71b9487bf56e9e", "output_hash": "0291810ef629db45b8dd44dbd2ffb05f6ec39daa8fe059b1ab9f757164cabc41"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "839a057109e056f55c03025a", "attempt_id": "974333e11e74", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:10:00.115228Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 6279, "message_tokens": 6193, "message_utf8_bytes": 30551, "cache_tensor_bytes": null, "serialized_payload_bytes": 30551, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.026957946829497814, "ttft_ms": 13.739082962274551, "decode_ms": 123.67087497841567, "receiver_prefill_ms": 818.2909580646083}, "total_wall_ms": 965.3328328859061, "ttft_ms": 13.739082962274551, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "3,875,883", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "bc7084590d7e4773b7e90ae4a9c7b9ead88f46214128f9ec5b71b9487bf56e9e", "output_hash": "a8243b862e2d82445ca3c93eec5e829e5d41cfa4d9abfba7b18d922fdebba636"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "2aa966e8671aa988eff389f2", "attempt_id": "770a02829d2c", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:10:01.100749Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 6289, "message_tokens": 6193, "message_utf8_bytes": 30551, "cache_tensor_bytes": null, "serialized_payload_bytes": 30551, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.026957946829497814, "ttft_ms": 14.044958050362766, "decode_ms": 83.03245902061462, "receiver_prefill_ms": 814.3432920332998}, "total_wall_ms": 921.181624988094, "ttft_ms": 14.044958050362766, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Rukmelri-6", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "bc7084590d7e4773b7e90ae4a9c7b9ead88f46214128f9ec5b71b9487bf56e9e", "output_hash": "2e6780effd3ce64b936cdd7bf5b52ba071dcd639cc48639e8f72d04cc6978a32"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c3abc9600e35af3521fbd186", "attempt_id": "4b46f3bae554", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:10:02.039544Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 6282, "message_tokens": 6193, "message_utf8_bytes": 30551, "cache_tensor_bytes": null, "serialized_payload_bytes": 30551, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.026957946829497814, "ttft_ms": 14.630790916271508, "decode_ms": 138.2669999729842, "receiver_prefill_ms": 844.2152080824599}, "total_wall_ms": 1006.842290982604, "ttft_ms": 14.630790916271508, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "30,455,553", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "bc7084590d7e4773b7e90ae4a9c7b9ead88f46214128f9ec5b71b9487bf56e9e", "output_hash": "5157f9aaa3594c1f0bd7a02ac6be1194a1ba8b10ab8bc810c8949c62ff6041f6"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "afb7cecf9a2751861521a20b", "attempt_id": "83755a92c2de", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:10:03.063537Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 6282, "message_tokens": 6193, "message_utf8_bytes": 30551, "cache_tensor_bytes": null, "serialized_payload_bytes": 30551, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.026957946829497814, "ttft_ms": 13.986374950036407, "decode_ms": 13.986125006340444, "receiver_prefill_ms": 811.5747079718858}, "total_wall_ms": 849.2370420135558, "ttft_ms": 13.986374950036407, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "bc7084590d7e4773b7e90ae4a9c7b9ead88f46214128f9ec5b71b9487bf56e9e", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a22300ccae72df65f95e59e1", "attempt_id": "bec6572b0b8e", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:10:03.929832Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 6284, "message_tokens": 6193, "message_utf8_bytes": 30551, "cache_tensor_bytes": null, "serialized_payload_bytes": 30551, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.026957946829497814, "ttft_ms": 14.020084054209292, "decode_ms": 13.764875009655952, "receiver_prefill_ms": 854.3699170695618}, "total_wall_ms": 891.9436669675633, "ttft_ms": 14.020084054209292, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "bc7084590d7e4773b7e90ae4a9c7b9ead88f46214128f9ec5b71b9487bf56e9e", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "dfb5ee4e75244cd432c0fdde", "attempt_id": "271f74e7cf86", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:10:30.188762Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1840}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 1952, "message_tokens": 1865, "message_utf8_bytes": 6372, "cache_tensor_bytes": null, "serialized_payload_bytes": 6372, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25349.24954199232, "ttft_ms": 13.527209055610001, "decode_ms": 93.16887496970594, "receiver_prefill_ms": 204.32779204566032}, "total_wall_ms": 25666.224709013477, "ttft_ms": 13.527209055610001, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "892,050", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "fb814772955e77dce8e5c2041c5879595b0aa8f64a88d209d6a0f00d176719fa", "output_hash": "eefb0cd4c83cfbb1178b00efbc639d605e7cf9215216943cfb9446639933dd9a"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "59e65fffbd9bf8ab34c6e536", "attempt_id": "6e711646de0a", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:10:30.524262Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1840}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 1955, "message_tokens": 1865, "message_utf8_bytes": 6372, "cache_tensor_bytes": null, "serialized_payload_bytes": 6372, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25349.24954199232, "ttft_ms": 13.432416948489845, "decode_ms": 133.8355840416625, "receiver_prefill_ms": 207.3804170358926}, "total_wall_ms": 25706.897041993216, "ttft_ms": 13.432416948489845, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "XX-43W-9874", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "fb814772955e77dce8e5c2041c5879595b0aa8f64a88d209d6a0f00d176719fa", "output_hash": "0ab6d0ab000bed72bda142906dbe114882d3e55d9e45808c6f2ae838a7201194"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "d2725dcfe47eb71524124d55", "attempt_id": "ca4dacb23da0", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:10:30.899128Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1840}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 1959, "message_tokens": 1865, "message_utf8_bytes": 6372, "cache_tensor_bytes": null, "serialized_payload_bytes": 6372, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25349.24954199232, "ttft_ms": 13.672083034180105, "decode_ms": 94.02429196052253, "receiver_prefill_ms": 207.1861249860376}, "total_wall_ms": 25667.09629201796, "ttft_ms": 13.672083034180105, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "931,295", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "fb814772955e77dce8e5c2041c5879595b0aa8f64a88d209d6a0f00d176719fa", "output_hash": "2a7fc9c11bafb9f078eeebdb6ef745c73bc462712e82c1adebdf9dc483e4d7ed"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "af4e5a13c1af9a44e127ec2d", "attempt_id": "0ebea5f74f6b", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:10:31.233956Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1840}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 1952, "message_tokens": 1865, "message_utf8_bytes": 6372, "cache_tensor_bytes": null, "serialized_payload_bytes": 6372, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25349.24954199232, "ttft_ms": 13.636709074489772, "decode_ms": 121.20254198089242, "receiver_prefill_ms": 212.13062503375113}, "total_wall_ms": 25699.278124957345, "ttft_ms": 13.636709074489772, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "3,418,930", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "fb814772955e77dce8e5c2041c5879595b0aa8f64a88d209d6a0f00d176719fa", "output_hash": "0c81bdc5a7cd40d6d4b0c52dc5fc252e6a9579c7b15fd2d5c7cfa4590ccab621"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "bca65390c527e56a0cacd1cc", "attempt_id": "7813e0abf1ce", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:10:31.601745Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1840}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 1962, "message_tokens": 1865, "message_utf8_bytes": 6372, "cache_tensor_bytes": null, "serialized_payload_bytes": 6372, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25349.24954199232, "ttft_ms": 13.799292035400867, "decode_ms": 80.74229105841368, "receiver_prefill_ms": 225.87316599674523}, "total_wall_ms": 25672.651959001087, "ttft_ms": 13.799292035400867, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Rukmelri-6", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "fb814772955e77dce8e5c2041c5879595b0aa8f64a88d209d6a0f00d176719fa", "output_hash": "2e6780effd3ce64b936cdd7bf5b52ba071dcd639cc48639e8f72d04cc6978a32"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "e8adb1f2364ed1514282c4b1", "attempt_id": "0112dfdf51c1", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:10:31.942355Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1840}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 1955, "message_tokens": 1865, "message_utf8_bytes": 6372, "cache_tensor_bytes": null, "serialized_payload_bytes": 6372, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25349.24954199232, "ttft_ms": 13.773082988336682, "decode_ms": 135.3177500423044, "receiver_prefill_ms": 232.88970801513642}, "total_wall_ms": 25734.141583903693, "ttft_ms": 13.773082988336682, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "20,544,825", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "fb814772955e77dce8e5c2041c5879595b0aa8f64a88d209d6a0f00d176719fa", "output_hash": "1b452395c99dcca7dd480d75e517d2127ade8ca1754f5241af3cff61d188786c"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "684fbe30826c3e547297dd01", "attempt_id": "b9660a3e7210", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:10:32.344196Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1840}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 1955, "message_tokens": 1865, "message_utf8_bytes": 6372, "cache_tensor_bytes": null, "serialized_payload_bytes": 6372, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25349.24954199232, "ttft_ms": 13.95712501835078, "decode_ms": 13.312833034433424, "receiver_prefill_ms": 219.354624976404}, "total_wall_ms": 25598.793583922088, "ttft_ms": 13.95712501835078, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "fb814772955e77dce8e5c2041c5879595b0aa8f64a88d209d6a0f00d176719fa", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "61081912abcb0fb3280a0f5a", "attempt_id": "9f41031a03f8", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:10:32.610850Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1840}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 1957, "message_tokens": 1865, "message_utf8_bytes": 6372, "cache_tensor_bytes": null, "serialized_payload_bytes": 6372, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25349.24954199232, "ttft_ms": 14.301458024419844, "decode_ms": 13.709584018215537, "receiver_prefill_ms": 238.91416599508375}, "total_wall_ms": 25619.17762493249, "ttft_ms": 14.301458024419844, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "fb814772955e77dce8e5c2041c5879595b0aa8f64a88d209d6a0f00d176719fa", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b2771019bdd1bcd1f428f64f", "attempt_id": "5e05a0f0a07d", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:10:55.783551Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 23}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 1013, "message_tokens": 922, "message_utf8_bytes": 2525, "cache_tensor_bytes": null, "serialized_payload_bytes": 2525, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 22885.207750019617, "ttft_ms": 13.43562500551343, "decode_ms": 13.777540996670723, "receiver_prefill_ms": 106.82633309625089}, "total_wall_ms": 23021.020000102, "ttft_ms": 13.43562500551343, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "aaaad4854733f81131cc2a023970957d9eb4802b404f757e5aff89d8f2b90e19", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "3a5b0b23b18e5ddd4830246b", "attempt_id": "28f79e24ae59", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:10:55.937298Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 23}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 1016, "message_tokens": 922, "message_utf8_bytes": 2525, "cache_tensor_bytes": null, "serialized_payload_bytes": 2525, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 22885.207750019617, "ttft_ms": 13.37387494277209, "decode_ms": 134.3983750557527, "receiver_prefill_ms": 107.0983330719173}, "total_wall_ms": 23141.760584083386, "ttft_ms": 13.37387494277209, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "XX-43W-9874", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "aaaad4854733f81131cc2a023970957d9eb4802b404f757e5aff89d8f2b90e19", "output_hash": "0ab6d0ab000bed72bda142906dbe114882d3e55d9e45808c6f2ae838a7201194"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "479ecda3a1538282c9a64d8e", "attempt_id": "d6dace6f7301", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:10:56.210806Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 23}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 1020, "message_tokens": 922, "message_utf8_bytes": 2525, "cache_tensor_bytes": null, "serialized_payload_bytes": 2525, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 22885.207750019617, "ttft_ms": 13.595666969195008, "decode_ms": 81.39662502799183, "receiver_prefill_ms": 107.3903749929741}, "total_wall_ms": 23089.304374996573, "ttft_ms": 13.595666969195008, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "931295", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "aaaad4854733f81131cc2a023970957d9eb4802b404f757e5aff89d8f2b90e19", "output_hash": "0291810ef629db45b8dd44dbd2ffb05f6ec39daa8fe059b1ab9f757164cabc41"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "eaf14d71e6809d2b85ef97e8", "attempt_id": "2a1d6b7d2a45", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:10:56.432303Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 23}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 1013, "message_tokens": 922, "message_utf8_bytes": 2525, "cache_tensor_bytes": null, "serialized_payload_bytes": 2525, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 22885.207750019617, "ttft_ms": 13.344166101887822, "decode_ms": 120.13770802877843, "receiver_prefill_ms": 106.29258397966623}, "total_wall_ms": 23126.636625034735, "ttft_ms": 13.344166101887822, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "3,875,883", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "aaaad4854733f81131cc2a023970957d9eb4802b404f757e5aff89d8f2b90e19", "output_hash": "a8243b862e2d82445ca3c93eec5e829e5d41cfa4d9abfba7b18d922fdebba636"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a10714dbaa570001638a4b4c", "attempt_id": "e235e5c13584", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:10:56.691379Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 23}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 1023, "message_tokens": 922, "message_utf8_bytes": 2525, "cache_tensor_bytes": null, "serialized_payload_bytes": 2525, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 22885.207750019617, "ttft_ms": 13.703417032957077, "decode_ms": 80.57741692755371, "receiver_prefill_ms": 117.9813330527395}, "total_wall_ms": 23099.12649996113, "ttft_ms": 13.703417032957077, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Project Talrukzen-5", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "aaaad4854733f81131cc2a023970957d9eb4802b404f757e5aff89d8f2b90e19", "output_hash": "7df0f16b0e459caba51aa78727eb6508fd7ce3f7ce3bd197e8a27bc183f5194e"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "84d9781476497e4be62c653d", "attempt_id": "fa3c3852792d", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:10:56.922837Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 23}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 1016, "message_tokens": 922, "message_utf8_bytes": 2525, "cache_tensor_bytes": null, "serialized_payload_bytes": 2525, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 22885.207750019617, "ttft_ms": 13.40979104861617, "decode_ms": 133.11154209077358, "receiver_prefill_ms": 106.50966700632125}, "total_wall_ms": 23139.986000023782, "ttft_ms": 13.40979104861617, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "22,451,303", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "aaaad4854733f81131cc2a023970957d9eb4802b404f757e5aff89d8f2b90e19", "output_hash": "15df1ace14b2dd753d8a5fb9e7b3ee4a0f9405b43e833f800830f4881f0d38b2"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "eb6bcaefdf5e1f2f1c285085", "attempt_id": "536f706f3ef7", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:10:57.196022Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 23}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 1016, "message_tokens": 922, "message_utf8_bytes": 2525, "cache_tensor_bytes": null, "serialized_payload_bytes": 2525, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 22885.207750019617, "ttft_ms": 13.340707984752953, "decode_ms": 13.126124977134168, "receiver_prefill_ms": 106.82854196056724}, "total_wall_ms": 23020.08420904167, "ttft_ms": 13.340707984752953, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "aaaad4854733f81131cc2a023970957d9eb4802b404f757e5aff89d8f2b90e19", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "83b8bc716e2ce58c81762383", "attempt_id": "ddaed2c1192d", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:10:57.348602Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 23}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 1018, "message_tokens": 922, "message_utf8_bytes": 2525, "cache_tensor_bytes": null, "serialized_payload_bytes": 2525, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 22885.207750019617, "ttft_ms": 13.387290993705392, "decode_ms": 13.40399996843189, "receiver_prefill_ms": 191.91954098641872}, "total_wall_ms": 23105.618541012518, "ttft_ms": 13.387290993705392, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "aaaad4854733f81131cc2a023970957d9eb4802b404f757e5aff89d8f2b90e19", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "07305bb4441db8c7ee3c8e55", "attempt_id": "3a8eaccd3c41", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:11:20.637687Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 2505, "message_tokens": 2415, "message_utf8_bytes": 6376, "cache_tensor_bytes": null, "serialized_payload_bytes": 6376, "side_information_bytes": 0, "fetched_bytes": 3507, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 23051.04995891452, "ttft_ms": 16.539958072826266, "decode_ms": 92.47687505558133, "receiver_prefill_ms": 842.7116659004241}, "total_wall_ms": 24011.045250925235, "ttft_ms": 16.539958072826266, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "892005", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "8eef720a14c2da3cab3ddf9361797bd89ace91393d97e62490bff2d6c2f5856f", "output_hash": "296832ce0f2ee0fb1dcf407ccce524ec73af796101cbe85b60c358235bf11a7e"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "448e60cac286de7bb187cf4e", "attempt_id": "9e8d4a399ca2", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:11:21.616791Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 2386, "message_tokens": 2293, "message_utf8_bytes": 6144, "cache_tensor_bytes": null, "serialized_payload_bytes": 6144, "side_information_bytes": 0, "fetched_bytes": 3281, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 23051.04995891452, "ttft_ms": 13.73620901722461, "decode_ms": 137.4157500686124, "receiver_prefill_ms": 262.54804199561477}, "total_wall_ms": 23475.969833903946, "ttft_ms": 13.73620901722461, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "XX-43W-9874", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "cb4bb6c758f1cd561c67a1a283f051867fe5c43c07b807bfa377b1275a0f8453", "output_hash": "0ab6d0ab000bed72bda142906dbe114882d3e55d9e45808c6f2ae838a7201194"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "41c7cf6872797b3ec07c2328", "attempt_id": "e3cfe21ed544", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:11:22.071379Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 2527, "message_tokens": 2430, "message_utf8_bytes": 6417, "cache_tensor_bytes": null, "serialized_payload_bytes": 6417, "side_information_bytes": 0, "fetched_bytes": 3545, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 23051.04995891452, "ttft_ms": 13.953416026197374, "decode_ms": 81.49266697000712, "receiver_prefill_ms": 282.0510419551283}, "total_wall_ms": 23432.001750916243, "ttft_ms": 13.953416026197374, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "931295", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "2d2eba9a2e57af7198e2235318d82d27d876ed2da991a5f9d5d54196c3c0d51e", "output_hash": "0291810ef629db45b8dd44dbd2ffb05f6ec39daa8fe059b1ab9f757164cabc41"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "15e70924ca1d2c9f3459a89a", "attempt_id": "bcf189ed3b99", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:11:22.470725Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 2515, "message_tokens": 2425, "message_utf8_bytes": 6392, "cache_tensor_bytes": null, "serialized_payload_bytes": 6392, "side_information_bytes": 0, "fetched_bytes": 3525, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 23051.04995891452, "ttft_ms": 14.130125055089593, "decode_ms": 125.28720800764859, "receiver_prefill_ms": 285.210374975577}, "total_wall_ms": 23479.042166960426, "ttft_ms": 14.130125055089593, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "3,875,883", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "f7acc8a71075f8b6f76b970c6a2b2e62fdbeede993c8bdae967422d8b7395378", "output_hash": "a8243b862e2d82445ca3c93eec5e829e5d41cfa4d9abfba7b18d922fdebba636"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "401041d331ecc0d80c68f8e8", "attempt_id": "79c9845515c6", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:11:22.916379Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 2417, "message_tokens": 2317, "message_utf8_bytes": 6223, "cache_tensor_bytes": null, "serialized_payload_bytes": 6223, "side_information_bytes": 0, "fetched_bytes": 3349, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 23051.04995891452, "ttft_ms": 14.060333021916449, "decode_ms": 82.70224998705089, "receiver_prefill_ms": 304.6272909268737}, "total_wall_ms": 23455.70787601173, "ttft_ms": 14.060333021916449, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Project Talrukzen-5", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "fcc9107e1f64ab43808a38fec503a557ec3468734dd3ca4a4a805e165a1d05cd", "output_hash": "7df0f16b0e459caba51aa78727eb6508fd7ce3f7ce3bd197e8a27bc183f5194e"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "8c81de9fbb76628a3d92c23a", "attempt_id": "98f182ecbdf6", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:11:23.338642Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 2524, "message_tokens": 2431, "message_utf8_bytes": 6419, "cache_tensor_bytes": null, "serialized_payload_bytes": 6419, "side_information_bytes": 0, "fetched_bytes": 3547, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 23051.04995891452, "ttft_ms": 13.956250040791929, "decode_ms": 137.60333298705518, "receiver_prefill_ms": 325.8884160313755}, "total_wall_ms": 23531.79720893968, "ttft_ms": 13.956250040791929, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "20,441,770", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "7836405e73f7560bb3c3830c1465d21a71eceeae12e7b2cee302d9e7a7a0db2b", "output_hash": "f3e6671d38bfe3834ed7dc7ac84ff0a3e4607d02c402c69579fc147364785f11"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "dc66c760de180901ade864c9", "attempt_id": "92ef2c286d2e", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:11:23.836482Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 2464, "message_tokens": 2371, "message_utf8_bytes": 6324, "cache_tensor_bytes": null, "serialized_payload_bytes": 6324, "side_information_bytes": 0, "fetched_bytes": 3435, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 23051.04995891452, "ttft_ms": 13.876667013391852, "decode_ms": 14.08987503964454, "receiver_prefill_ms": 308.3404579665512}, "total_wall_ms": 23390.76658396516, "ttft_ms": 13.876667013391852, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "cdfd4924d395fdbf3de41c053c691fb191210db3576ed637d9d6c24613a94c24", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "9b52ad94afda515b73fc6172", "attempt_id": "8a416cfbabb4", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:11:24.194016Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 2514, "message_tokens": 2419, "message_utf8_bytes": 6350, "cache_tensor_bytes": null, "serialized_payload_bytes": 6350, "side_information_bytes": 0, "fetched_bytes": 3471, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 23051.04995891452, "ttft_ms": 13.877290999516845, "decode_ms": 13.750958954915404, "receiver_prefill_ms": 325.1212500035763}, "total_wall_ms": 23407.225875882432, "ttft_ms": 13.877290999516845, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "c74dddd1f3b73a1bfc83eccfff96068505a0521b470c8ae154412472b58e897f", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "22614ada4a6f2148a1ccf0b4", "attempt_id": "089771ed33af", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:11:26.743373Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 234049536}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 44, "message_tokens": null, "message_utf8_bytes": 30497, "cache_tensor_bytes": 234049536, "serialized_payload_bytes": 234049536, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 2175.7419579662383, "ttft_ms": 450.59999998193234, "decode_ms": 141.59291703253984, "receiver_prefill_ms": 0.0}, "total_wall_ms": 2861.6725000320002, "ttft_ms": 450.59999998193234, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5,782,865", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "696bd2cb8a823e73776453d0fb303898c309a6b51df05f83968e15ea29c836ab"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "d1d08cf3a5337a7d0e58a277", "attempt_id": "55949eedb643", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:11:27.447658Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 234049536}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 30497, "cache_tensor_bytes": 234049536, "serialized_payload_bytes": 234049536, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 2175.7419579662383, "ttft_ms": 82.58016698528081, "decode_ms": 151.175000006333, "receiver_prefill_ms": 0.0}, "total_wall_ms": 2414.807458058931, "ttft_ms": 82.58016698528081, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "XX-43W-9874", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "0ab6d0ab000bed72bda142906dbe114882d3e55d9e45808c6f2ae838a7201194"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "36d3edb4624190303f9728ee", "attempt_id": "ac8fc3dde36b", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:11:27.705300Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 234049536}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 51, "message_tokens": null, "message_utf8_bytes": 30497, "cache_tensor_bytes": 234049536, "serialized_payload_bytes": 234049536, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 2175.7419579662383, "ttft_ms": 123.8270839676261, "decode_ms": 105.86675000377, "receiver_prefill_ms": 0.0}, "total_wall_ms": 2409.8808738635853, "ttft_ms": 123.8270839676261, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "931,295", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "2a7fc9c11bafb9f078eeebdb6ef745c73bc462712e82c1adebdf9dc483e4d7ed"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "16bb3de3cdcbc394f76c7f53", "attempt_id": "6cbdf95b9bdd", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:11:27.957071Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 234049536}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 44, "message_tokens": null, "message_utf8_bytes": 30497, "cache_tensor_bytes": 234049536, "serialized_payload_bytes": 234049536, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 2175.7419579662383, "ttft_ms": 49.966542050242424, "decode_ms": 135.6238330481574, "receiver_prefill_ms": 0.0}, "total_wall_ms": 2364.884207956493, "ttft_ms": 49.966542050242424, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "3,418,930", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "0c81bdc5a7cd40d6d4b0c52dc5fc252e6a9579c7b15fd2d5c7cfa4590ccab621"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "6ce8be1ec337a7d56e814beb", "attempt_id": "21c86424850e", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:11:28.163584Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 234049536}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 54, "message_tokens": null, "message_utf8_bytes": 30497, "cache_tensor_bytes": 234049536, "serialized_payload_bytes": 234049536, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 2175.7419579662383, "ttft_ms": 166.6190419346094, "decode_ms": 91.53091709595174, "receiver_prefill_ms": 0.0}, "total_wall_ms": 2438.3321249624714, "ttft_ms": 166.6190419346094, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Rukmelri-6", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "2e6780effd3ce64b936cdd7bf5b52ba071dcd639cc48639e8f72d04cc6978a32"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "790ee42fcd3c62a69e88baa8", "attempt_id": "95e8e69ed114", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:11:28.443390Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 234049536}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 30497, "cache_tensor_bytes": 234049536, "serialized_payload_bytes": 234049536, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 2175.7419579662383, "ttft_ms": 42.71391697693616, "decode_ms": 120.0127499178052, "receiver_prefill_ms": 0.0}, "total_wall_ms": 2342.983082984574, "ttft_ms": 42.71391697693616, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "10421705", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "d87d2d53f8f676159d5d45b8f3f51545551bb9b20f2270f2888cfcda09aa8f94"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "70ccf010dd22fc0d484845a1", "attempt_id": "c5a5c82c349e", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:11:28.628469Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 234049536}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 30497, "cache_tensor_bytes": 234049536, "serialized_payload_bytes": 234049536, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 2175.7419579662383, "ttft_ms": 42.81424998771399, "decode_ms": 14.8896659957245, "receiver_prefill_ms": 0.0}, "total_wall_ms": 2237.9913330078125, "ttft_ms": 42.81424998771399, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c2bf0288c7131e5de5aab7e3", "attempt_id": "93a64ab521ed", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:11:28.708393Z", "task_family": "synthetic", "document_id": "syn_03010_8000", "document_hash": "8beb6214b362911c46f5db753e36cc61c5444a7d8ea3c267dd1346c90c9d36f7", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03010_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 234049536}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5787, "query_tokens": 49, "message_tokens": null, "message_utf8_bytes": 30497, "cache_tensor_bytes": 234049536, "serialized_payload_bytes": 234049536, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 2175.7419579662383, "ttft_ms": 47.351041925139725, "decode_ms": 15.224125003442168, "receiver_prefill_ms": 0.0}, "total_wall_ms": 2242.901291931048, "ttft_ms": 47.351041925139725, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Talrukzen-5", "bogus_amount": "5,735,925", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03010_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "4e53ce46fcf756315a1fe7eb", "attempt_id": "1d38d08cdfe5", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:11:29.758680Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 2367, "message_tokens": 2283, "message_utf8_bytes": 8557, "cache_tensor_bytes": null, "serialized_payload_bytes": 8557, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.04866707604378462, "ttft_ms": 13.443542062304914, "decode_ms": 94.12470797542483, "receiver_prefill_ms": 272.78037497308105}, "total_wall_ms": 384.6754591213539, "ttft_ms": 13.443542062304914, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8631552", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "21f0f06f2ac70c1f810dffe5439619e911faabde878e44971abd26434090f4b3", "output_hash": "147d02c0587a77b3b53f0424c8d3d1454aaf2639ce9fdaf69af20db33363e849"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b2f643b11f825c7606e91b99", "attempt_id": "dee737542df0", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:11:30.157237Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 2370, "message_tokens": 2283, "message_utf8_bytes": 8557, "cache_tensor_bytes": null, "serialized_payload_bytes": 8557, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.04866707604378462, "ttft_ms": 13.594000018201768, "decode_ms": 134.32537496555597, "receiver_prefill_ms": 261.1754999961704}, "total_wall_ms": 413.2610420929268, "ttft_ms": 13.594000018201768, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "DG-40E-2777", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "21f0f06f2ac70c1f810dffe5439619e911faabde878e44971abd26434090f4b3", "output_hash": "16a7e3f4a7cc4a9b66f86c2cdcfad3b35daf1da199d9406e0627c8af2d8d73f3"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "81d77985d6d81a97c3e8501f", "attempt_id": "6030c96d4ce0", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:11:30.584277Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 2377, "message_tokens": 2283, "message_utf8_bytes": 8557, "cache_tensor_bytes": null, "serialized_payload_bytes": 8557, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.04866707604378462, "ttft_ms": 13.39879201259464, "decode_ms": 93.81074993871152, "receiver_prefill_ms": 261.2536660162732}, "total_wall_ms": 372.02245905064046, "ttft_ms": 13.39879201259464, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "6602245", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "21f0f06f2ac70c1f810dffe5439619e911faabde878e44971abd26434090f4b3", "output_hash": "d117fa8197988f8d1e4fabc6c4cdaab9516018e54553b57c63b875b3de82b16f"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "27b562265958ab4bc2c6e85e", "attempt_id": "c56c213e8a0b", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:11:30.970045Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 2370, "message_tokens": 2283, "message_utf8_bytes": 8557, "cache_tensor_bytes": null, "serialized_payload_bytes": 8557, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.04866707604378462, "ttft_ms": 13.430166989564896, "decode_ms": 80.60004201252013, "receiver_prefill_ms": 260.6837920611724}, "total_wall_ms": 358.3374171284959, "ttft_ms": 13.430166989564896, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "466045", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "21f0f06f2ac70c1f810dffe5439619e911faabde878e44971abd26434090f4b3", "output_hash": "b87a05c1686ed24576a21ca1da9c3967d1980c731d8c73962a30d77a67087945"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "44f84b1a821f2e30ca8be630", "attempt_id": "5d31ac6d3b2c", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:11:31.342563Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 2380, "message_tokens": 2283, "message_utf8_bytes": 8557, "cache_tensor_bytes": null, "serialized_payload_bytes": 8557, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.04866707604378462, "ttft_ms": 13.2884579943493, "decode_ms": 80.15041693579406, "receiver_prefill_ms": 262.2525420738384}, "total_wall_ms": 359.3069590860978, "ttft_ms": 13.2884579943493, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Felaldfel-9", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "21f0f06f2ac70c1f810dffe5439619e911faabde878e44971abd26434090f4b3", "output_hash": "09ab1578c99909d3c22a0c8042977ad718c576e7b4cce4e50abc683b5f7d9a64"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "43decb5b00923e200d1445ad", "attempt_id": "64e473da90ee", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:11:31.715364Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 2372, "message_tokens": 2283, "message_utf8_bytes": 8557, "cache_tensor_bytes": null, "serialized_payload_bytes": 8557, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.04866707604378462, "ttft_ms": 13.28716694843024, "decode_ms": 94.33554101269692, "receiver_prefill_ms": 261.99304196052253}, "total_wall_ms": 373.2582920929417, "ttft_ms": 13.28716694843024, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "2108155", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "21f0f06f2ac70c1f810dffe5439619e911faabde878e44971abd26434090f4b3", "output_hash": "97d7580e36d137e0001a925fc01a2b872098ac407a5737a1de282c78fb69debf"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "81c13e4d30e9085f7ef69c60", "attempt_id": "702bffec3bf8", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:11:32.102803Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 2374, "message_tokens": 2283, "message_utf8_bytes": 8557, "cache_tensor_bytes": null, "serialized_payload_bytes": 8557, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.04866707604378462, "ttft_ms": 13.576083001680672, "decode_ms": 13.429625076241791, "receiver_prefill_ms": 261.67854096274823}, "total_wall_ms": 292.2835841309279, "ttft_ms": 13.576083001680672, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "21f0f06f2ac70c1f810dffe5439619e911faabde878e44971abd26434090f4b3", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a503a5e94b03205a10b406b2", "attempt_id": "ff69233fbed9", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:11:32.408381Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 2374, "message_tokens": 2283, "message_utf8_bytes": 8557, "cache_tensor_bytes": null, "serialized_payload_bytes": 8557, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.04866707604378462, "ttft_ms": 13.319709105417132, "decode_ms": 13.334708055481315, "receiver_prefill_ms": 261.140082962811}, "total_wall_ms": 291.3758751237765, "ttft_ms": 13.319709105417132, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "21f0f06f2ac70c1f810dffe5439619e911faabde878e44971abd26434090f4b3", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "9f4c0a89666e1ded92053c5f", "attempt_id": "420814063bbb", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:11:46.684574Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 1123, "message_tokens": 1037, "message_utf8_bytes": 2568, "cache_tensor_bytes": null, "serialized_payload_bytes": 2568, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13970.546083059162, "ttft_ms": 13.216791907325387, "decode_ms": 119.83908305410296, "receiver_prefill_ms": 121.8877910869196}, "total_wall_ms": 14227.898000157438, "ttft_ms": 13.216791907325387, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8,631,552", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "bfc3c472795eb43389ede3b4facd3345dad5ada691623b956c03d9f2e85a1104", "output_hash": "cecdbfc66b98662e79ca5deba218a2b11cf8d9e18b8f3d91c3032acf0c41fc29"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "1106ef8181d4ebefca878887", "attempt_id": "2fb789e27661", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:11:46.956779Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 1126, "message_tokens": 1037, "message_utf8_bytes": 2568, "cache_tensor_bytes": null, "serialized_payload_bytes": 2568, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13970.546083059162, "ttft_ms": 13.283708016388118, "decode_ms": 133.42004199512303, "receiver_prefill_ms": 122.12470907252282}, "total_wall_ms": 14241.131999064237, "ttft_ms": 13.283708016388118, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "DG-40E-2777", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "bfc3c472795eb43389ede3b4facd3345dad5ada691623b956c03d9f2e85a1104", "output_hash": "16a7e3f4a7cc4a9b66f86c2cdcfad3b35daf1da199d9406e0627c8af2d8d73f3"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a862118c6befd54499fdadac", "attempt_id": "a64113b994bd", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:11:47.241500Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 1133, "message_tokens": 1037, "message_utf8_bytes": 2568, "cache_tensor_bytes": null, "serialized_payload_bytes": 2568, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13970.546083059162, "ttft_ms": 13.3081249659881, "decode_ms": 119.10833290312439, "receiver_prefill_ms": 121.44458305556327}, "total_wall_ms": 14226.18879203219, "ttft_ms": 13.3081249659881, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "6,602,245", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "bfc3c472795eb43389ede3b4facd3345dad5ada691623b956c03d9f2e85a1104", "output_hash": "f121947078d464ab91a83739861b98c336a10c001704e2df84cc2b041bb8f4de"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "5a474eff23f1f1b93d8e6784", "attempt_id": "76b0ed8ef1f5", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:11:47.511404Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 1126, "message_tokens": 1037, "message_utf8_bytes": 2568, "cache_tensor_bytes": null, "serialized_payload_bytes": 2568, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13970.546083059162, "ttft_ms": 13.339999946765602, "decode_ms": 79.69974994193763, "receiver_prefill_ms": 121.15699995774776}, "total_wall_ms": 14186.471374006942, "ttft_ms": 13.339999946765602, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "466045", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "bfc3c472795eb43389ede3b4facd3345dad5ada691623b956c03d9f2e85a1104", "output_hash": "b87a05c1686ed24576a21ca1da9c3967d1980c731d8c73962a30d77a67087945"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c2d075db5548af9d0ea35bf1", "attempt_id": "f90cda38a28b", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:11:47.741921Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 1136, "message_tokens": 1037, "message_utf8_bytes": 2568, "cache_tensor_bytes": null, "serialized_payload_bytes": 2568, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13970.546083059162, "ttft_ms": 13.461375026963651, "decode_ms": 79.69375001266599, "receiver_prefill_ms": 121.41120806336403}, "total_wall_ms": 14186.856458080001, "ttft_ms": 13.461375026963651, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Felaldfel-9", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "bfc3c472795eb43389ede3b4facd3345dad5ada691623b956c03d9f2e85a1104", "output_hash": "09ab1578c99909d3c22a0c8042977ad718c576e7b4cce4e50abc683b5f7d9a64"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "ab13ef65be00b1c658c7debb", "attempt_id": "88f14d289cd6", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:11:47.971998Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 1128, "message_tokens": 1037, "message_utf8_bytes": 2568, "cache_tensor_bytes": null, "serialized_payload_bytes": 2568, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13970.546083059162, "ttft_ms": 13.426791992969811, "decode_ms": 132.13808392174542, "receiver_prefill_ms": 121.05129100382328}, "total_wall_ms": 14238.955333130434, "ttft_ms": 13.426791992969811, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "17,302,947", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "bfc3c472795eb43389ede3b4facd3345dad5ada691623b956c03d9f2e85a1104", "output_hash": "56bfc8152b15fc690beb7e2381c5c17e39dd0db2397c85b7453b2255edda09c1"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "97b3dcc8ae784029cb63e015", "attempt_id": "8675ebc48cbd", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:11:48.254471Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 1130, "message_tokens": 1037, "message_utf8_bytes": 2568, "cache_tensor_bytes": null, "serialized_payload_bytes": 2568, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13970.546083059162, "ttft_ms": 13.208458898589015, "decode_ms": 13.277542078867555, "receiver_prefill_ms": 121.20008294004947}, "total_wall_ms": 14119.949583080597, "ttft_ms": 13.208458898589015, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "bfc3c472795eb43389ede3b4facd3345dad5ada691623b956c03d9f2e85a1104", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "3349d10db8742ef5cd4c763a", "attempt_id": "9eb0eaf3e527", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:11:48.418427Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 1130, "message_tokens": 1037, "message_utf8_bytes": 2568, "cache_tensor_bytes": null, "serialized_payload_bytes": 2568, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13970.546083059162, "ttft_ms": 13.415541965514421, "decode_ms": 13.635416980832815, "receiver_prefill_ms": 121.01883394643664}, "total_wall_ms": 14120.359042077325, "ttft_ms": 13.415541965514421, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "bfc3c472795eb43389ede3b4facd3345dad5ada691623b956c03d9f2e85a1104", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "9d9b079def455d420f349dcc", "attempt_id": "e77c89cfed6f", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:12:02.569283Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 18}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 775, "message_tokens": 686, "message_utf8_bytes": 1851, "cache_tensor_bytes": null, "serialized_payload_bytes": 1851, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13986.53275007382, "ttft_ms": 13.3004579693079, "decode_ms": 93.42225000727922, "receiver_prefill_ms": 92.71737502422184}, "total_wall_ms": 14187.503791996278, "ttft_ms": 13.3004579693079, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5690750", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "074b3df8cba3c20890fe938e1126a2929cf9b26e2ea04ccb779bf4caeb79ac4a", "output_hash": "7be636c93357eea3aed60ce58daa563e90a051598f8161ba3719afca87db6194"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "2e8aa436beeb70fae0f2eb2a", "attempt_id": "5e862c6e4e52", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:12:02.785094Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 18}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 778, "message_tokens": 686, "message_utf8_bytes": 1851, "cache_tensor_bytes": null, "serialized_payload_bytes": 1851, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13986.53275007382, "ttft_ms": 13.181542046368122, "decode_ms": 132.4103750521317, "receiver_prefill_ms": 92.37020800355822}, "total_wall_ms": 14225.970000145026, "ttft_ms": 13.181542046368122, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "DG-40E-2777", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "074b3df8cba3c20890fe938e1126a2929cf9b26e2ea04ccb779bf4caeb79ac4a", "output_hash": "16a7e3f4a7cc4a9b66f86c2cdcfad3b35daf1da199d9406e0627c8af2d8d73f3"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "d0fe9e0f3e08deb865d60959", "attempt_id": "d0bc91fbaa27", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:12:03.039030Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 18}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 785, "message_tokens": 686, "message_utf8_bytes": 1851, "cache_tensor_bytes": null, "serialized_payload_bytes": 1851, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13986.53275007382, "ttft_ms": 13.359916978515685, "decode_ms": 92.66383305657655, "receiver_prefill_ms": 92.00545900966972}, "total_wall_ms": 14185.95208413899, "ttft_ms": 13.359916978515685, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5690750", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "074b3df8cba3c20890fe938e1126a2929cf9b26e2ea04ccb779bf4caeb79ac4a", "output_hash": "7be636c93357eea3aed60ce58daa563e90a051598f8161ba3719afca87db6194"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "fc594f4feb92ff8127782aa8", "attempt_id": "36de50308220", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:12:03.252361Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 18}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 778, "message_tokens": 686, "message_utf8_bytes": 1851, "cache_tensor_bytes": null, "serialized_payload_bytes": 1851, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13986.53275007382, "ttft_ms": 13.147833989933133, "decode_ms": 13.426917023025453, "receiver_prefill_ms": 91.69870789628476}, "total_wall_ms": 14106.221500085667, "ttft_ms": 13.147833989933133, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "074b3df8cba3c20890fe938e1126a2929cf9b26e2ea04ccb779bf4caeb79ac4a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "87b47d90dd20766f5295fec2", "attempt_id": "0c0fb7489ee2", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:12:03.386167Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 18}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 788, "message_tokens": 686, "message_utf8_bytes": 1851, "cache_tensor_bytes": null, "serialized_payload_bytes": 1851, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13986.53275007382, "ttft_ms": 13.203249895013869, "decode_ms": 13.36829096544534, "receiver_prefill_ms": 92.40600001066923}, "total_wall_ms": 14107.000958058052, "ttft_ms": 13.203249895013869, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "074b3df8cba3c20890fe938e1126a2929cf9b26e2ea04ccb779bf4caeb79ac4a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "3b151625de70c5c192b7b221", "attempt_id": "18ae27d2fa91", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:12:03.520580Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 18}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 780, "message_tokens": 686, "message_utf8_bytes": 1851, "cache_tensor_bytes": null, "serialized_payload_bytes": 1851, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13986.53275007382, "ttft_ms": 13.174833031371236, "decode_ms": 91.92849998362362, "receiver_prefill_ms": 91.6915830457583}, "total_wall_ms": 14184.706875123084, "ttft_ms": 13.174833031371236, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5690750", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "074b3df8cba3c20890fe938e1126a2929cf9b26e2ea04ccb779bf4caeb79ac4a", "output_hash": "7be636c93357eea3aed60ce58daa563e90a051598f8161ba3719afca87db6194"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "0911cc8ffb51cda4d01026c5", "attempt_id": "16bf4a476ec9", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:12:03.732516Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 18}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 782, "message_tokens": 686, "message_utf8_bytes": 1851, "cache_tensor_bytes": null, "serialized_payload_bytes": 1851, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13986.53275007382, "ttft_ms": 13.323624967597425, "decode_ms": 13.481708010658622, "receiver_prefill_ms": 91.75270795822144}, "total_wall_ms": 14106.442084070295, "ttft_ms": 13.323624967597425, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "074b3df8cba3c20890fe938e1126a2929cf9b26e2ea04ccb779bf4caeb79ac4a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "edcee303ca305ba28415b16d", "attempt_id": "993170f9e993", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:12:03.866213Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 18}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 782, "message_tokens": 686, "message_utf8_bytes": 1851, "cache_tensor_bytes": null, "serialized_payload_bytes": 1851, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13986.53275007382, "ttft_ms": 13.252165983431041, "decode_ms": 13.120874995365739, "receiver_prefill_ms": 91.35624999180436}, "total_wall_ms": 14105.617125052959, "ttft_ms": 13.252165983431041, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "074b3df8cba3c20890fe938e1126a2929cf9b26e2ea04ccb779bf4caeb79ac4a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "4d2ff60344f55a6a21f29ca1", "attempt_id": "744233b58028", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:12:18.030697Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 752, "message_tokens": 664, "message_utf8_bytes": 1745, "cache_tensor_bytes": null, "serialized_payload_bytes": 1745, "side_information_bytes": 0, "fetched_bytes": 825, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14031.338209053501, "ttft_ms": 13.286834000609815, "decode_ms": 13.276334037072957, "receiver_prefill_ms": 81.5679159713909}, "total_wall_ms": 14141.003874945454, "ttft_ms": 13.286834000609815, "input_usage": {"lookups": 3}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "41b2cd593230b983a4231c8d0e5e03000fdb457b7d1b92a96fd363826fef8364", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "eb9c2d3b3a3a5a71a48fe405", "attempt_id": "7737adab6f07", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:12:18.154967Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 749, "message_tokens": 658, "message_utf8_bytes": 1702, "cache_tensor_bytes": null, "serialized_payload_bytes": 1702, "side_information_bytes": 0, "fetched_bytes": 771, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14031.338209053501, "ttft_ms": 13.17316701170057, "decode_ms": 132.18612503260374, "receiver_prefill_ms": 80.82849998027086}, "total_wall_ms": 14258.95695912186, "ttft_ms": 13.17316701170057, "input_usage": {"lookups": 3}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "DG-40E-2777", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "732854ad03266ca17dcf0a2360513e8f1f218a174d4fbada20df0e36c80fa140", "output_hash": "16a7e3f4a7cc4a9b66f86c2cdcfad3b35daf1da199d9406e0627c8af2d8d73f3"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "56ede31ee5ee8952680006bc", "attempt_id": "140157a31619", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:12:18.396725Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 762, "message_tokens": 664, "message_utf8_bytes": 1745, "cache_tensor_bytes": null, "serialized_payload_bytes": 1745, "side_information_bytes": 0, "fetched_bytes": 825, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14031.338209053501, "ttft_ms": 13.065916951745749, "decode_ms": 13.330040965229273, "receiver_prefill_ms": 81.93137508351356}, "total_wall_ms": 14141.06925111264, "ttft_ms": 13.065916951745749, "input_usage": {"lookups": 3}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "41b2cd593230b983a4231c8d0e5e03000fdb457b7d1b92a96fd363826fef8364", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "d79bd2c2265deeac3a1c99fa", "attempt_id": "243d550184a9", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:12:18.520637Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 755, "message_tokens": 664, "message_utf8_bytes": 1745, "cache_tensor_bytes": null, "serialized_payload_bytes": 1745, "side_information_bytes": 0, "fetched_bytes": 825, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14031.338209053501, "ttft_ms": 13.035833020694554, "decode_ms": 13.26041598804295, "receiver_prefill_ms": 81.5633749589324}, "total_wall_ms": 14140.57162613608, "ttft_ms": 13.035833020694554, "input_usage": {"lookups": 3}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "41b2cd593230b983a4231c8d0e5e03000fdb457b7d1b92a96fd363826fef8364", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "2d7a9897490753d559ad1c68", "attempt_id": "56ed4b2ca29a", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:12:18.644238Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 765, "message_tokens": 664, "message_utf8_bytes": 1745, "cache_tensor_bytes": null, "serialized_payload_bytes": 1745, "side_information_bytes": 0, "fetched_bytes": 825, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14031.338209053501, "ttft_ms": 13.205958064645529, "decode_ms": 13.084292062558234, "receiver_prefill_ms": 81.59912494011223}, "total_wall_ms": 14140.654250979424, "ttft_ms": 13.205958064645529, "input_usage": {"lookups": 3}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "41b2cd593230b983a4231c8d0e5e03000fdb457b7d1b92a96fd363826fef8364", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "06e5c81bf072a14715fd3224", "attempt_id": "44ef1d665670", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:12:18.767483Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 757, "message_tokens": 664, "message_utf8_bytes": 1745, "cache_tensor_bytes": null, "serialized_payload_bytes": 1745, "side_information_bytes": 0, "fetched_bytes": 825, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14031.338209053501, "ttft_ms": 13.129874947480857, "decode_ms": 106.5416659694165, "receiver_prefill_ms": 81.30512502975762}, "total_wall_ms": 14233.880584128201, "ttft_ms": 13.129874947480857, "input_usage": {"lookups": 3}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "18763670", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "41b2cd593230b983a4231c8d0e5e03000fdb457b7d1b92a96fd363826fef8364", "output_hash": "9871914ddf6f1b5e53c1857fb98454c144ef43c95d8256a8b73fb75433f5e750"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "347563f09f6df20ce7b3d017", "attempt_id": "cd7bafd52482", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:12:18.984234Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 760, "message_tokens": 665, "message_utf8_bytes": 1766, "cache_tensor_bytes": null, "serialized_payload_bytes": 1766, "side_information_bytes": 0, "fetched_bytes": 831, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14031.338209053501, "ttft_ms": 13.160333968698978, "decode_ms": 13.041917001828551, "receiver_prefill_ms": 82.66462502069771}, "total_wall_ms": 14141.965918126516, "ttft_ms": 13.160333968698978, "input_usage": {"lookups": 3}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "e254fa98bb723d8a51e21c36995794ad6d541e347ba610e44bfc906b998c6b54", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "14c2136db126a5b482dca43d", "attempt_id": "7d4812eda532", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:12:19.108750Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 759, "message_tokens": 664, "message_utf8_bytes": 1745, "cache_tensor_bytes": null, "serialized_payload_bytes": 1745, "side_information_bytes": 0, "fetched_bytes": 825, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14031.338209053501, "ttft_ms": 13.236667029559612, "decode_ms": 13.07745894882828, "receiver_prefill_ms": 81.17895794566721}, "total_wall_ms": 14140.263459063135, "ttft_ms": 13.236667029559612, "input_usage": {"lookups": 3}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "41b2cd593230b983a4231c8d0e5e03000fdb457b7d1b92a96fd363826fef8364", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "7c85b0ee175740f8104bfa3b", "attempt_id": "91a87b81dd48", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:12:19.504629Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 87035904}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 42, "message_tokens": null, "message_utf8_bytes": 8531, "cache_tensor_bytes": 87035904, "serialized_payload_bytes": 87035904, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 273.0322909774259, "ttft_ms": 31.81458404287696, "decode_ms": 125.05799997597933, "receiver_prefill_ms": 0.0}, "total_wall_ms": 434.1535409912467, "ttft_ms": 31.81458404287696, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5,690,750", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "8f9fd54847380148ac21931745ca6e07a6e1f79a19682e0c182f511a5fbb45de"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "3cffeb7f8ef735901f87d293", "attempt_id": "d669c76ca377", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:12:19.680209Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 87035904}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 45, "message_tokens": null, "message_utf8_bytes": 8531, "cache_tensor_bytes": 87035904, "serialized_payload_bytes": 87035904, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 273.0322909774259, "ttft_ms": 28.731458936817944, "decode_ms": 138.0917920032516, "receiver_prefill_ms": 0.0}, "total_wall_ms": 443.30670696217567, "ttft_ms": 28.731458936817944, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "XX-71M-6213", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "66ef6c5da682b38795e9029723bbd71cdde2c4dec70cf1fae6efe62ef243f10a"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "80505bccf2143d052b6e90a5", "attempt_id": "29053ebc9427", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:12:19.864595Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 87035904}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 52, "message_tokens": null, "message_utf8_bytes": 8531, "cache_tensor_bytes": 87035904, "serialized_payload_bytes": 87035904, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 273.0322909774259, "ttft_ms": 28.56416697613895, "decode_ms": 83.39520904701203, "receiver_prefill_ms": 0.0}, "total_wall_ms": 388.5134159354493, "ttft_ms": 28.56416697613895, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "587410", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b2d10a78599ca584d9d082e01bc8678a94ed6593dee465579d37518a0f4c7afd"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "bb60399769bdb33e1056485d", "attempt_id": "7411423feb1a", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:12:19.994430Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 87035904}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 45, "message_tokens": null, "message_utf8_bytes": 8531, "cache_tensor_bytes": 87035904, "serialized_payload_bytes": 87035904, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 273.0322909774259, "ttft_ms": 28.469334007240832, "decode_ms": 82.72766706068069, "receiver_prefill_ms": 0.0}, "total_wall_ms": 387.47645798139274, "ttft_ms": 28.469334007240832, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "587410", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b2d10a78599ca584d9d082e01bc8678a94ed6593dee465579d37518a0f4c7afd"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c2cb7bbba5a2ebc46cc584c8", "attempt_id": "f2996ced8d1f", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:12:20.123220Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 87035904}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 55, "message_tokens": null, "message_utf8_bytes": 8531, "cache_tensor_bytes": 87035904, "serialized_payload_bytes": 87035904, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 273.0322909774259, "ttft_ms": 27.81833300832659, "decode_ms": 82.88308302871883, "receiver_prefill_ms": 0.0}, "total_wall_ms": 387.6056659501046, "ttft_ms": 27.81833300832659, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Lunsol-15", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "2d7097497adac2e578d6d73b20fc698db21401ab49da96bec88f50124ccc2dc7"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "2426ac9c947bdebabe053d01", "attempt_id": "e2297df61f44", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:12:20.251788Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 87035904}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 8531, "cache_tensor_bytes": 87035904, "serialized_payload_bytes": 87035904, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 273.0322909774259, "ttft_ms": 27.541875024326146, "decode_ms": 96.36712505016476, "receiver_prefill_ms": 0.0}, "total_wall_ms": 400.97295795567334, "ttft_ms": 27.541875024326146, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1492247", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "a6557643a97d27daf76dd9f5c85264be3c45be9fb4749efe86e4f1b8e7dcfc21"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "339d77293d278bbb1b3f3793", "attempt_id": "e8b9c9bf4525", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:12:20.397541Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 87035904}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 49, "message_tokens": null, "message_utf8_bytes": 8531, "cache_tensor_bytes": 87035904, "serialized_payload_bytes": 87035904, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 273.0322909774259, "ttft_ms": 28.013500035740435, "decode_ms": 13.86945799458772, "receiver_prefill_ms": 0.0}, "total_wall_ms": 318.8231660751626, "ttft_ms": 28.013500035740435, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c0ddc0c23eb0001123bb4cdd", "attempt_id": "e4de37a0973d", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:12:20.457208Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 87035904}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 49, "message_tokens": null, "message_utf8_bytes": 8531, "cache_tensor_bytes": 87035904, "serialized_payload_bytes": 87035904, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 273.0322909774259, "ttft_ms": 27.76266692671925, "decode_ms": 14.24495899118483, "receiver_prefill_ms": 0.0}, "total_wall_ms": 318.73641593847424, "ttft_ms": 27.76266692671925, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "24a31806f29b212dbc9138f8", "attempt_id": "eb1bb7d7f2fe", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:12:20.517329Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 2481, "message_tokens": 2397, "message_utf8_bytes": 9061, "cache_tensor_bytes": null, "serialized_payload_bytes": 9061, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.02391706220805645, "ttft_ms": 13.715083012357354, "decode_ms": 94.05679197516292, "receiver_prefill_ms": 270.36929107271135}, "total_wall_ms": 381.9000830408186, "ttft_ms": 13.715083012357354, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8631552", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S4"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ca586f4bee09dc8b9141d20a36e0ccc5e9e72a56cab6dab031fcaedb699595a3", "output_hash": "147d02c0587a77b3b53f0424c8d3d1454aaf2639ce9fdaf69af20db33363e849"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b86e357dc0b86bc5fc716fb3", "attempt_id": "37674af8ad04", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:12:20.913292Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 2484, "message_tokens": 2397, "message_utf8_bytes": 9061, "cache_tensor_bytes": null, "serialized_payload_bytes": 9061, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.02391706220805645, "ttft_ms": 13.694874942302704, "decode_ms": 134.80750005692244, "receiver_prefill_ms": 275.19841701723635}, "total_wall_ms": 427.4217920610681, "ttft_ms": 13.694874942302704, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "DG-40E-2777", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S4"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ca586f4bee09dc8b9141d20a36e0ccc5e9e72a56cab6dab031fcaedb699595a3", "output_hash": "16a7e3f4a7cc4a9b66f86c2cdcfad3b35daf1da199d9406e0627c8af2d8d73f3"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "dc9af3a97e28d34e00de86da", "attempt_id": "4849a2ca0e46", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:12:21.354745Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 2491, "message_tokens": 2397, "message_utf8_bytes": 9061, "cache_tensor_bytes": null, "serialized_payload_bytes": 9061, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.02391706220805645, "ttft_ms": 13.555416953749955, "decode_ms": 94.57137493882328, "receiver_prefill_ms": 277.01670909300447}, "total_wall_ms": 388.89320904854685, "ttft_ms": 13.555416953749955, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "6602245", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S4"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ca586f4bee09dc8b9141d20a36e0ccc5e9e72a56cab6dab031fcaedb699595a3", "output_hash": "d117fa8197988f8d1e4fabc6c4cdaab9516018e54553b57c63b875b3de82b16f"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "8e7ddde521b4ee4d2612bb28", "attempt_id": "64d404468615", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:12:21.757872Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 2484, "message_tokens": 2397, "message_utf8_bytes": 9061, "cache_tensor_bytes": null, "serialized_payload_bytes": 9061, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.02391706220805645, "ttft_ms": 13.52866692468524, "decode_ms": 80.98062500357628, "receiver_prefill_ms": 278.2064579660073}, "total_wall_ms": 376.37179205194116, "ttft_ms": 13.52866692468524, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "587410", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S4"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ca586f4bee09dc8b9141d20a36e0ccc5e9e72a56cab6dab031fcaedb699595a3", "output_hash": "b2d10a78599ca584d9d082e01bc8678a94ed6593dee465579d37518a0f4c7afd"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "54d45ba08bf9f69d9c241b3d", "attempt_id": "bf3f5f57cebe", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:12:22.148219Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 2494, "message_tokens": 2397, "message_utf8_bytes": 9061, "cache_tensor_bytes": null, "serialized_payload_bytes": 9061, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.02391706220805645, "ttft_ms": 13.743292074650526, "decode_ms": 81.66566700674593, "receiver_prefill_ms": 281.7261249292642}, "total_wall_ms": 380.882001016289, "ttft_ms": 13.743292074650526, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Felaldfel-9", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S4"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ca586f4bee09dc8b9141d20a36e0ccc5e9e72a56cab6dab031fcaedb699595a3", "output_hash": "09ab1578c99909d3c22a0c8042977ad718c576e7b4cce4e50abc683b5f7d9a64"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "ebb2e95883c28d20be0ebf09", "attempt_id": "4e34df58f71b", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:12:22.543713Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 2486, "message_tokens": 2397, "message_utf8_bytes": 9061, "cache_tensor_bytes": null, "serialized_payload_bytes": 9061, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.02391706220805645, "ttft_ms": 13.999792048707604, "decode_ms": 95.02987505402416, "receiver_prefill_ms": 296.00708303041756}, "total_wall_ms": 408.72937499079853, "ttft_ms": 13.999792048707604, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "3404058", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S4"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ca586f4bee09dc8b9141d20a36e0ccc5e9e72a56cab6dab031fcaedb699595a3", "output_hash": "b68aa8e6b7e35f2899553c114d40014ca9aa9697ef523bf4f5dcb358d145645a"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "5f3bac83b09d15153e99a1f2", "attempt_id": "486329e7e89c", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:12:22.966726Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 2488, "message_tokens": 2397, "message_utf8_bytes": 9061, "cache_tensor_bytes": null, "serialized_payload_bytes": 9061, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.02391706220805645, "ttft_ms": 13.70420795865357, "decode_ms": 13.654750073328614, "receiver_prefill_ms": 287.8029169514775}, "total_wall_ms": 318.9183750655502, "ttft_ms": 13.70420795865357, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S4"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ca586f4bee09dc8b9141d20a36e0ccc5e9e72a56cab6dab031fcaedb699595a3", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "5dbec5ada40a38a69cab7ca9", "attempt_id": "9092e31d284c", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:12:23.299876Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 2488, "message_tokens": 2397, "message_utf8_bytes": 9061, "cache_tensor_bytes": null, "serialized_payload_bytes": 9061, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.02391706220805645, "ttft_ms": 13.713167048990726, "decode_ms": 13.418041984550655, "receiver_prefill_ms": 296.3320000562817}, "total_wall_ms": 327.27204205002636, "ttft_ms": 13.713167048990726, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S4"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ca586f4bee09dc8b9141d20a36e0ccc5e9e72a56cab6dab031fcaedb699595a3", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a84f755a9f8154b762f1f9d2", "attempt_id": "c6ad5761d7a2", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:12:48.463499Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1835}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 1947, "message_tokens": 1862, "message_utf8_bytes": 6088, "cache_tensor_bytes": null, "serialized_payload_bytes": 6088, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24822.187584009953, "ttft_ms": 13.334334013052285, "decode_ms": 94.23083299770951, "receiver_prefill_ms": 204.24520899541676}, "total_wall_ms": 25137.685750960372, "ttft_ms": 13.334334013052285, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8631552", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S4"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "85538681a747ec9f00d89d17bcfc8f88b9ff5b214337e7f5905cf874ec5ab733", "output_hash": "147d02c0587a77b3b53f0424c8d3d1454aaf2639ce9fdaf69af20db33363e849"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "d8804c04de15b52739a37d0a", "attempt_id": "3ca677b77d27", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:12:48.793590Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1835}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 1950, "message_tokens": 1862, "message_utf8_bytes": 6088, "cache_tensor_bytes": null, "serialized_payload_bytes": 6088, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24822.187584009953, "ttft_ms": 13.387665967456996, "decode_ms": 132.8689579386264, "receiver_prefill_ms": 203.73195793945342}, "total_wall_ms": 25175.177876022644, "ttft_ms": 13.387665967456996, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "DG-40E-2777", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S4"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "85538681a747ec9f00d89d17bcfc8f88b9ff5b214337e7f5905cf874ec5ab733", "output_hash": "16a7e3f4a7cc4a9b66f86c2cdcfad3b35daf1da199d9406e0627c8af2d8d73f3"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "1ca2577a3838d950e1e8cb10", "attempt_id": "266dcec0f233", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:12:49.160842Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1835}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 1957, "message_tokens": 1862, "message_utf8_bytes": 6088, "cache_tensor_bytes": null, "serialized_payload_bytes": 6088, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24822.187584009953, "ttft_ms": 13.389374944381416, "decode_ms": 93.9415420871228, "receiver_prefill_ms": 206.38258394319564}, "total_wall_ms": 25138.838542974554, "ttft_ms": 13.389374944381416, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5899340", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S4"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "85538681a747ec9f00d89d17bcfc8f88b9ff5b214337e7f5905cf874ec5ab733", "output_hash": "c920323a9a02b1afb8619cf11bdde427f7cd1ea3289be69bd62932985895d717"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b4579d5540b72dfe34e087c8", "attempt_id": "748c0afcfc15", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:12:49.494017Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1835}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 1950, "message_tokens": 1862, "message_utf8_bytes": 6088, "cache_tensor_bytes": null, "serialized_payload_bytes": 6088, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24822.187584009953, "ttft_ms": 13.454416999593377, "decode_ms": 81.01862506009638, "receiver_prefill_ms": 209.05258296988904}, "total_wall_ms": 25128.65762610454, "ttft_ms": 13.454416999593377, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "466045", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S4"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "85538681a747ec9f00d89d17bcfc8f88b9ff5b214337e7f5905cf874ec5ab733", "output_hash": "b87a05c1686ed24576a21ca1da9c3967d1980c731d8c73962a30d77a67087945"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "58be700a37055e00e4a10797", "attempt_id": "f83fbc76b3d0", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:12:49.814717Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1835}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 1960, "message_tokens": 1862, "message_utf8_bytes": 6088, "cache_tensor_bytes": null, "serialized_payload_bytes": 6088, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24822.187584009953, "ttft_ms": 13.5163749800995, "decode_ms": 81.1953330412507, "receiver_prefill_ms": 228.3203329425305}, "total_wall_ms": 25148.206418030895, "ttft_ms": 13.5163749800995, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Lunsol-15", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S4"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "85538681a747ec9f00d89d17bcfc8f88b9ff5b214337e7f5905cf874ec5ab733", "output_hash": "2d7097497adac2e578d6d73b20fc698db21401ab49da96bec88f50124ccc2dc7"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "676c85f3372ed5a111c26e03", "attempt_id": "ad3faada579a", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:12:50.155198Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1835}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 1952, "message_tokens": 1862, "message_utf8_bytes": 6088, "cache_tensor_bytes": null, "serialized_payload_bytes": 6088, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24822.187584009953, "ttft_ms": 13.784999959170818, "decode_ms": 108.68591698817909, "receiver_prefill_ms": 225.95012502279133}, "total_wall_ms": 25173.578875954263, "ttft_ms": 13.784999959170818, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "30942535", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S4"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "85538681a747ec9f00d89d17bcfc8f88b9ff5b214337e7f5905cf874ec5ab733", "output_hash": "895413f51f35d738fa58f98b42f815915c6182f08fc7901e2f6e6062726ee755"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "30fcb04c3cf38241c64695a7", "attempt_id": "41b9e2b8af7c", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:12:50.520867Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1835}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 1954, "message_tokens": 1862, "message_utf8_bytes": 6088, "cache_tensor_bytes": null, "serialized_payload_bytes": 6088, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24822.187584009953, "ttft_ms": 13.669500011019409, "decode_ms": 13.668374973349273, "receiver_prefill_ms": 219.75137491244823}, "total_wall_ms": 25072.21583405044, "ttft_ms": 13.669500011019409, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S4"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "85538681a747ec9f00d89d17bcfc8f88b9ff5b214337e7f5905cf874ec5ab733", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "018505128c8281196bd00125", "attempt_id": "1e128eb09b36", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:12:50.785147Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1835}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 1954, "message_tokens": 1862, "message_utf8_bytes": 6088, "cache_tensor_bytes": null, "serialized_payload_bytes": 6088, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24822.187584009953, "ttft_ms": 14.344083960168064, "decode_ms": 13.707167003303766, "receiver_prefill_ms": 239.69808407127857}, "total_wall_ms": 25092.91075100191, "ttft_ms": 14.344083960168064, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S4"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "85538681a747ec9f00d89d17bcfc8f88b9ff5b214337e7f5905cf874ec5ab733", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "43c9a1135636ed8d2e05737a", "attempt_id": "efd330209af4", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:13:18.667301Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 31}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 1263, "message_tokens": 1174, "message_utf8_bytes": 3116, "cache_tensor_bytes": null, "serialized_payload_bytes": 3116, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27596.992334001698, "ttft_ms": 13.393500004895031, "decode_ms": 93.33154105115682, "receiver_prefill_ms": 131.1515000415966}, "total_wall_ms": 27836.781251011416, "ttft_ms": 13.393500004895031, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8631552", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S4"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "679b4f933de9bfbee7e944dd2dc5287c507dcf208686064436f6cf6af14c1b8c", "output_hash": "147d02c0587a77b3b53f0424c8d3d1454aaf2639ce9fdaf69af20db33363e849"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "90022c8091dd3da06070c2d6", "attempt_id": "c005b892451a", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:13:18.921735Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 31}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 1266, "message_tokens": 1174, "message_utf8_bytes": 3116, "cache_tensor_bytes": null, "serialized_payload_bytes": 3116, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27596.992334001698, "ttft_ms": 14.529332984238863, "decode_ms": 139.19708400499076, "receiver_prefill_ms": 161.33345803245902}, "total_wall_ms": 27914.038334041834, "ttft_ms": 14.529332984238863, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "DG-40E-2777", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S4"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "679b4f933de9bfbee7e944dd2dc5287c507dcf208686064436f6cf6af14c1b8c", "output_hash": "16a7e3f4a7cc4a9b66f86c2cdcfad3b35daf1da199d9406e0627c8af2d8d73f3"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "065b8127ee47acc7035967f4", "attempt_id": "bf8f5669f5d6", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:13:19.254916Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 31}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 1273, "message_tokens": 1174, "message_utf8_bytes": 3116, "cache_tensor_bytes": null, "serialized_payload_bytes": 3116, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27596.992334001698, "ttft_ms": 56.32129102014005, "decode_ms": 97.83995791804045, "receiver_prefill_ms": 2238.3372500771657}, "total_wall_ms": 29991.453833994456, "ttft_ms": 56.32129102014005, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "6602245", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S4"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "679b4f933de9bfbee7e944dd2dc5287c507dcf208686064436f6cf6af14c1b8c", "output_hash": "d117fa8197988f8d1e4fabc6c4cdaab9516018e54553b57c63b875b3de82b16f"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "41fe0f6dff8e0ae1e6cb5661", "attempt_id": "5b20cf2ddd25", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:13:21.663681Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 31}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 1266, "message_tokens": 1174, "message_utf8_bytes": 3116, "cache_tensor_bytes": null, "serialized_payload_bytes": 3116, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27596.992334001698, "ttft_ms": 14.428416965529323, "decode_ms": 14.207375003024936, "receiver_prefill_ms": 158.47174997907132}, "total_wall_ms": 27786.397416959517, "ttft_ms": 14.428416965529323, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S4"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "679b4f933de9bfbee7e944dd2dc5287c507dcf208686064436f6cf6af14c1b8c", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "4013e83e85a7e2a4f8f0c63e", "attempt_id": "f97bfee03572", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:13:21.869943Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 31}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 1276, "message_tokens": 1174, "message_utf8_bytes": 3116, "cache_tensor_bytes": null, "serialized_payload_bytes": 3116, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27596.992334001698, "ttft_ms": 13.718375004827976, "decode_ms": 85.65579203423113, "receiver_prefill_ms": 144.45091597735882}, "total_wall_ms": 27843.10329204891, "ttft_ms": 13.718375004827976, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Felaldfel-9", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S4"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "679b4f933de9bfbee7e944dd2dc5287c507dcf208686064436f6cf6af14c1b8c", "output_hash": "09ab1578c99909d3c22a0c8042977ad718c576e7b4cce4e50abc683b5f7d9a64"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "66f526a53f386b2f32ebfc54", "attempt_id": "110fb0bd3677", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:13:22.131462Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 31}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 1268, "message_tokens": 1174, "message_utf8_bytes": 3116, "cache_tensor_bytes": null, "serialized_payload_bytes": 3116, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27596.992334001698, "ttft_ms": 13.3896250044927, "decode_ms": 95.35900002811104, "receiver_prefill_ms": 136.36220793705434}, "total_wall_ms": 27844.00637599174, "ttft_ms": 13.3896250044927, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "2454355", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S4"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "679b4f933de9bfbee7e944dd2dc5287c507dcf208686064436f6cf6af14c1b8c", "output_hash": "7b4e158da46994d5aa334f14c3bfc05093c7846139bbb75e522665b545a9fe09"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "155b3870dc52ee420a427fd4", "attempt_id": "8b484775253b", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:13:22.393094Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 31}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 1270, "message_tokens": 1174, "message_utf8_bytes": 3116, "cache_tensor_bytes": null, "serialized_payload_bytes": 3116, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27596.992334001698, "ttft_ms": 13.912875088863075, "decode_ms": 13.788540964014828, "receiver_prefill_ms": 245.55604194756597}, "total_wall_ms": 27872.426001005806, "ttft_ms": 13.912875088863075, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S4"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "679b4f933de9bfbee7e944dd2dc5287c507dcf208686064436f6cf6af14c1b8c", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "aded2ec6a9b6bcfcc6c9312f", "attempt_id": "e4aa2b676768", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:13:22.682597Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 31}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 1270, "message_tokens": 1174, "message_utf8_bytes": 3116, "cache_tensor_bytes": null, "serialized_payload_bytes": 3116, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27596.992334001698, "ttft_ms": 14.189583947882056, "decode_ms": 14.066166942939162, "receiver_prefill_ms": 130.9999160002917}, "total_wall_ms": 27758.500209078193, "ttft_ms": 14.189583947882056, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S4"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "679b4f933de9bfbee7e944dd2dc5287c507dcf208686064436f6cf6af14c1b8c", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "d3f751431892c44d33aee108", "attempt_id": "2580f4b5bfd3", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:13:51.127995Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 2603, "message_tokens": 2515, "message_utf8_bytes": 6675, "cache_tensor_bytes": null, "serialized_payload_bytes": 6675, "side_information_bytes": 0, "fetched_bytes": 3448, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 28265.4844999779, "ttft_ms": 15.257709077559412, "decode_ms": 100.98237497732043, "receiver_prefill_ms": 521.8893330311403}, "total_wall_ms": 28908.284083008766, "ttft_ms": 15.257709077559412, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8631552", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S4"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "02cbe5b0ee8b7f307b970dd631dcecb37f57cc89d64cd68ad4c4ed3f3b3168db", "output_hash": "147d02c0587a77b3b53f0424c8d3d1454aaf2639ce9fdaf69af20db33363e849"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "f2f66aba78c7dc578ae67f27", "attempt_id": "4e370b6ca0b5", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:13:51.788381Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 2524, "message_tokens": 2433, "message_utf8_bytes": 6465, "cache_tensor_bytes": null, "serialized_payload_bytes": 6465, "side_information_bytes": 0, "fetched_bytes": 3220, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 28265.4844999779, "ttft_ms": 14.109584037214518, "decode_ms": 137.53779197577387, "receiver_prefill_ms": 381.3215830596164}, "total_wall_ms": 28802.88649990689, "ttft_ms": 14.109584037214518, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "DG-40E-2777", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S4"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "b1c7be3311bc228152549bae22f83a03f87e7202a22e1e78f729eae09556e6a9", "output_hash": "16a7e3f4a7cc4a9b66f86c2cdcfad3b35daf1da199d9406e0627c8af2d8d73f3"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "5ebc516461eec3ccaf0a2e65", "attempt_id": "d25e833e30d5", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:13:52.342710Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 2635, "message_tokens": 2537, "message_utf8_bytes": 6645, "cache_tensor_bytes": null, "serialized_payload_bytes": 6645, "side_information_bytes": 0, "fetched_bytes": 3426, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 28265.4844999779, "ttft_ms": 13.989083003252745, "decode_ms": 94.39458302222192, "receiver_prefill_ms": 287.0250000851229}, "total_wall_ms": 28664.453790988773, "ttft_ms": 13.989083003252745, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "6602245", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S4"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "5d7cbd89c4eb30f45578d16b5b3f0d79e6716a276270ec01f94255ed32a766e6", "output_hash": "d117fa8197988f8d1e4fabc6c4cdaab9516018e54553b57c63b875b3de82b16f"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "63971ca1c5948f99f6c6d39c", "attempt_id": "251d61f60d15", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:13:52.757345Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 2658, "message_tokens": 2567, "message_utf8_bytes": 6704, "cache_tensor_bytes": null, "serialized_payload_bytes": 6704, "side_information_bytes": 0, "fetched_bytes": 3483, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 28265.4844999779, "ttft_ms": 14.087958028540015, "decode_ms": 82.24358293227851, "receiver_prefill_ms": 304.2949589435011}, "total_wall_ms": 28670.321749988943, "ttft_ms": 14.087958028540015, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "466045", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S4"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "e6ae511ee08ef653c18bebcf57091addc95cf98095c451b5bb0af7e039d00260", "output_hash": "b87a05c1686ed24576a21ca1da9c3967d1980c731d8c73962a30d77a67087945"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "227120fa8f17a44bc8dec6bf", "attempt_id": "f9d5fe278edc", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:13:53.176999Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 2564, "message_tokens": 2463, "message_utf8_bytes": 6563, "cache_tensor_bytes": null, "serialized_payload_bytes": 6563, "side_information_bytes": 0, "fetched_bytes": 3351, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 28265.4844999779, "ttft_ms": 13.849458075128496, "decode_ms": 82.27033307775855, "receiver_prefill_ms": 335.59708402026445}, "total_wall_ms": 28700.684209004976, "ttft_ms": 13.849458075128496, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Lunsol-15", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S4"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "2a3f92a890f7a3da19f07c752fc5f1d265edc4e21b700c7f8c57b49306ec4b5c", "output_hash": "2d7097497adac2e578d6d73b20fc698db21401ab49da96bec88f50124ccc2dc7"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "d17179ba89a7f05388aa01b3", "attempt_id": "a387758ff5a6", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:13:53.626886Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 2630, "message_tokens": 2537, "message_utf8_bytes": 6645, "cache_tensor_bytes": null, "serialized_payload_bytes": 6645, "side_information_bytes": 0, "fetched_bytes": 3426, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 28265.4844999779, "ttft_ms": 13.638291973620653, "decode_ms": 136.3406250020489, "receiver_prefill_ms": 348.42637507244945}, "total_wall_ms": 28767.250000033528, "ttft_ms": 13.638291973620653, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "33,723,385", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S4"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "5d7cbd89c4eb30f45578d16b5b3f0d79e6716a276270ec01f94255ed32a766e6", "output_hash": "0ce1a2a2ed58cc6a96b99db4b17dca38004e9b1d426ca733509d9c04a4840f0b"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "df0fde236a471be7ef128eca", "attempt_id": "acce80493e2a", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:13:54.143168Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 2581, "message_tokens": 2486, "message_utf8_bytes": 6577, "cache_tensor_bytes": null, "serialized_payload_bytes": 6577, "side_information_bytes": 0, "fetched_bytes": 3320, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 28265.4844999779, "ttft_ms": 13.735625077970326, "decode_ms": 13.582624960690737, "receiver_prefill_ms": 298.59029094222933}, "total_wall_ms": 28594.700374873355, "ttft_ms": 13.735625077970326, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S4"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "02072ed6755a5b14909d7068ddd484c96d6cf62d4bc9678ee5d00544459859a7", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "8184d7ad5f8ab4b39e0c81c5", "attempt_id": "53bb1e86ee90", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:13:54.486471Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 2560, "message_tokens": 2465, "message_utf8_bytes": 6606, "cache_tensor_bytes": null, "serialized_payload_bytes": 6606, "side_information_bytes": 0, "fetched_bytes": 3379, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 28265.4844999779, "ttft_ms": 13.583375024609268, "decode_ms": 14.450832968577743, "receiver_prefill_ms": 323.47862503957003}, "total_wall_ms": 28620.368667063303, "ttft_ms": 13.583375024609268, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S4"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "554f624f72092ea153643a86aad019614497f49ab34ada331da7bcc3ef1f8f2a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "2b7d5de3c1fe446b1a71fdda", "attempt_id": "5578e31c265f", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:13:55.281323Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 94113792}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 42, "message_tokens": null, "message_utf8_bytes": 9007, "cache_tensor_bytes": 94113792, "serialized_payload_bytes": 94113792, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 425.43100006878376, "ttft_ms": 135.74862491805106, "decode_ms": 185.14449999202043, "receiver_prefill_ms": 0.0}, "total_wall_ms": 751.8761660903692, "ttft_ms": 135.74862491805106, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5,690,750", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S4"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "8f9fd54847380148ac21931745ca6e07a6e1f79a19682e0c182f511a5fbb45de"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "6ea1e35c41f4b2737ffbe7ed", "attempt_id": "fb51746a83af", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:13:55.625794Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 94113792}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 45, "message_tokens": null, "message_utf8_bytes": 9007, "cache_tensor_bytes": 94113792, "serialized_payload_bytes": 94113792, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 425.43100006878376, "ttft_ms": 29.853125102818012, "decode_ms": 139.77895805146545, "receiver_prefill_ms": 0.0}, "total_wall_ms": 612.1090410742909, "ttft_ms": 29.853125102818012, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "YW-85X-2214", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S4"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "498166dac36b6cd44be048aa87aa47e5a6fd656ba6d16daf6ffe39e336ab898d"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "f5d763907e6f8cd2c8de99c8", "attempt_id": "a9ff07eb22ed", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:13:55.827071Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 94113792}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 52, "message_tokens": null, "message_utf8_bytes": 9007, "cache_tensor_bytes": 94113792, "serialized_payload_bytes": 94113792, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 425.43100006878376, "ttft_ms": 30.10366705711931, "decode_ms": 126.13400002010167, "receiver_prefill_ms": 0.0}, "total_wall_ms": 586.5970000158995, "ttft_ms": 30.10366705711931, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5,899,340", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S4"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "cdcb6f9099e6e651a99b4424bdf64b76bf516969970c3c7ba61297500761eb4f"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "81392ae26654d2139a330e85", "attempt_id": "594ddedfa2da", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:13:56.002653Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 94113792}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 45, "message_tokens": null, "message_utf8_bytes": 9007, "cache_tensor_bytes": 94113792, "serialized_payload_bytes": 94113792, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 425.43100006878376, "ttft_ms": 29.831124935299158, "decode_ms": 97.53641695715487, "receiver_prefill_ms": 0.0}, "total_wall_ms": 555.9120421530679, "ttft_ms": 29.831124935299158, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "466,045", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S4"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "4a66e3c0960c26d2429d36e0c7f9d4c2192499734aa4ec831aea9100579be90f"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "794d28067a1a35b64fc036af", "attempt_id": "c3619173a8e1", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:13:56.147559Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 94113792}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 55, "message_tokens": null, "message_utf8_bytes": 9007, "cache_tensor_bytes": 94113792, "serialized_payload_bytes": 94113792, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 425.43100006878376, "ttft_ms": 29.067041934467852, "decode_ms": 83.80575000774115, "receiver_prefill_ms": 0.0}, "total_wall_ms": 541.7955840239301, "ttft_ms": 29.067041934467852, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Lunsol-15", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S4"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "2d7097497adac2e578d6d73b20fc698db21401ab49da96bec88f50124ccc2dc7"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "12d7035626e553bf1451309b", "attempt_id": "fb670b8d88bb", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:13:56.278519Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 94113792}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 9007, "cache_tensor_bytes": 94113792, "serialized_payload_bytes": 94113792, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 425.43100006878376, "ttft_ms": 28.417166089639068, "decode_ms": 110.82341696601361, "receiver_prefill_ms": 0.0}, "total_wall_ms": 567.9389161523432, "ttft_ms": 28.417166089639068, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "14042475", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S4"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "e356ce5cde97dcf37b63c711bc82d1758905f4865ec9081c5de48e579bc13823"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c8b77a963a9069b6b3b2d5fc", "attempt_id": "5507253b885f", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:13:56.435433Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 94113792}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 49, "message_tokens": null, "message_utf8_bytes": 9007, "cache_tensor_bytes": 94113792, "serialized_payload_bytes": 94113792, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 425.43100006878376, "ttft_ms": 28.964957920834422, "decode_ms": 14.087583986110985, "receiver_prefill_ms": 0.0}, "total_wall_ms": 471.70995909255, "ttft_ms": 28.964957920834422, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S4"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "9b5de1fde315b57ea3fd4a99", "attempt_id": "71e1b3c2e2d8", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:13:56.495726Z", "task_family": "synthetic", "document_id": "syn_03012_2000", "document_hash": "03e441e68d54e1af036f8ff905dfd6cbbccead62636279c1a8e179352ab08945", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03012_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 94113792}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1927, "query_tokens": 49, "message_tokens": null, "message_utf8_bytes": 9007, "cache_tensor_bytes": 94113792, "serialized_payload_bytes": 94113792, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 425.43100006878376, "ttft_ms": 28.893416048958898, "decode_ms": 13.87483300641179, "receiver_prefill_ms": 0.0}, "total_wall_ms": 471.3650420308113, "ttft_ms": 28.893416048958898, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Banyminov-7", "bogus_amount": "5,486,385", "source": "S4"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03012_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a21d8dbca7b90b17a56551f2", "attempt_id": "0997f862597c", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:13:57.505602Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 6255, "message_tokens": 6168, "message_utf8_bytes": 30401, "cache_tensor_bytes": null, "serialized_payload_bytes": 30401, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.3718750085681677, "ttft_ms": 13.883333071134984, "decode_ms": 122.99720803275704, "receiver_prefill_ms": 744.5959579199553}, "total_wall_ms": 897.3177500301972, "ttft_ms": 13.883333071134984, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "9,519,700", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "62261975ebeb2813b01d83fb28ed2a9136b1887f5bc73f7b24a07373d3386788", "output_hash": "f31f51fd5fb1986ddf80fc727e98ae3da80e1335676b76bd1bc1660a441d2eb2"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "46988e404cfb6e75689721c6", "attempt_id": "3e53d19b5ff4", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:13:58.419485Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 6256, "message_tokens": 6168, "message_utf8_bytes": 30401, "cache_tensor_bytes": null, "serialized_payload_bytes": 30401, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.3718750085681677, "ttft_ms": 13.72191694099456, "decode_ms": 137.17137498315424, "receiver_prefill_ms": 732.2878750273958}, "total_wall_ms": 893.0456669768319, "ttft_ms": 13.72191694099456, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "YA-85D-7885", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "62261975ebeb2813b01d83fb28ed2a9136b1887f5bc73f7b24a07373d3386788", "output_hash": "d7c446b684af4f3a0aa8481919c405aba8a334272947b087f62b1b6c80b3f946"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b4c1687e9af08d5d85d8e1cc", "attempt_id": "3bb045ddcc2a", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:13:59.329249Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 6261, "message_tokens": 6168, "message_utf8_bytes": 30401, "cache_tensor_bytes": null, "serialized_payload_bytes": 30401, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.3718750085681677, "ttft_ms": 14.249958912841976, "decode_ms": 97.1429169876501, "receiver_prefill_ms": 734.1096249874681}, "total_wall_ms": 855.337416054681, "ttft_ms": 14.249958912841976, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7782250", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "62261975ebeb2813b01d83fb28ed2a9136b1887f5bc73f7b24a07373d3386788", "output_hash": "35ed445e80874ee97060da399132c68930141ac0b1db960634d510ba5e921356"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "d9ed5236f8ea0aa3bee3d04a", "attempt_id": "ce5f7e0104a4", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:14:00.200752Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 6254, "message_tokens": 6168, "message_utf8_bytes": 30401, "cache_tensor_bytes": null, "serialized_payload_bytes": 30401, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.3718750085681677, "ttft_ms": 13.635459006763995, "decode_ms": 124.69366705045104, "receiver_prefill_ms": 733.7213329738006}, "total_wall_ms": 882.0715829497203, "ttft_ms": 13.635459006763995, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8,846,545", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "62261975ebeb2813b01d83fb28ed2a9136b1887f5bc73f7b24a07373d3386788", "output_hash": "77d7beaf301ce085d0b52f1ffccc70fe340d883e4f7677ef3ec61ef496cd7d5d"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "3eca9a87f63fc9c5906ed5df", "attempt_id": "fffbcd35d9e3", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:14:01.099451Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 6263, "message_tokens": 6168, "message_utf8_bytes": 30401, "cache_tensor_bytes": null, "serialized_payload_bytes": 30401, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.3718750085681677, "ttft_ms": 13.893250026740134, "decode_ms": 82.53679191693664, "receiver_prefill_ms": 741.3536250824109}, "total_wall_ms": 847.739625023678, "ttft_ms": 13.893250026740134, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Lithzenban-5", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "62261975ebeb2813b01d83fb28ed2a9136b1887f5bc73f7b24a07373d3386788", "output_hash": "cc5e85bb9beafdd57dcb7f11c04a12aea0492c44343bf9633324f1d50e9c4b6f"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "f067caf58561ecbc1eb9cb89", "attempt_id": "b6a6173da6a6", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:14:01.963712Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 6257, "message_tokens": 6168, "message_utf8_bytes": 30401, "cache_tensor_bytes": null, "serialized_payload_bytes": 30401, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.3718750085681677, "ttft_ms": 14.011207967996597, "decode_ms": 96.67004097718745, "receiver_prefill_ms": 762.748459004797}, "total_wall_ms": 883.2839580718428, "ttft_ms": 14.011207967996597, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4744458", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "62261975ebeb2813b01d83fb28ed2a9136b1887f5bc73f7b24a07373d3386788", "output_hash": "1fe5e7977423ae1379893c8626f0755f2444e23be99040fecd7bf2564b3b2c77"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "48a6e375a23bb7b3a595ff6e", "attempt_id": "936cb42281f4", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:14:02.863219Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 6257, "message_tokens": 6168, "message_utf8_bytes": 30401, "cache_tensor_bytes": null, "serialized_payload_bytes": 30401, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.3718750085681677, "ttft_ms": 13.7655840953812, "decode_ms": 13.994000037200749, "receiver_prefill_ms": 777.259707916528}, "total_wall_ms": 814.9329170119017, "ttft_ms": 13.7655840953812, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "62261975ebeb2813b01d83fb28ed2a9136b1887f5bc73f7b24a07373d3386788", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "89225abfcd9b119af4c46802", "attempt_id": "f90ab15a7e22", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:14:03.694760Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 6259, "message_tokens": 6168, "message_utf8_bytes": 30401, "cache_tensor_bytes": null, "serialized_payload_bytes": 30401, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.3718750085681677, "ttft_ms": 14.082583016715944, "decode_ms": 14.110542018897831, "receiver_prefill_ms": 811.291457968764}, "total_wall_ms": 849.5384170673788, "ttft_ms": 14.082583016715944, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "62261975ebeb2813b01d83fb28ed2a9136b1887f5bc73f7b24a07373d3386788", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c313a43736eeb6c75d04effd", "attempt_id": "443852ea5b0f", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:14:19.138834Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 1126, "message_tokens": 1037, "message_utf8_bytes": 3330, "cache_tensor_bytes": null, "serialized_payload_bytes": 3330, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14577.936624991708, "ttft_ms": 13.61745793838054, "decode_ms": 122.03895801212639, "receiver_prefill_ms": 121.98062508832663}, "total_wall_ms": 14843.34324998781, "ttft_ms": 13.61745793838054, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "9,519,700", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "e756d3601086b09c617d9e1191e356bbef034617cd9082ff9baec27a864a08f2", "output_hash": "f31f51fd5fb1986ddf80fc727e98ae3da80e1335676b76bd1bc1660a441d2eb2"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "6f23fbf7bcb059f5e6d26fce", "attempt_id": "22ccfd6c5a82", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:14:19.421779Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 1127, "message_tokens": 1037, "message_utf8_bytes": 3330, "cache_tensor_bytes": null, "serialized_payload_bytes": 3330, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14577.936624991708, "ttft_ms": 13.540250016376376, "decode_ms": 134.21366701368243, "receiver_prefill_ms": 121.45333306398243}, "total_wall_ms": 14849.13729201071, "ttft_ms": 13.540250016376376, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "YA-85D-7885", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "e756d3601086b09c617d9e1191e356bbef034617cd9082ff9baec27a864a08f2", "output_hash": "d7c446b684af4f3a0aa8481919c405aba8a334272947b087f62b1b6c80b3f946"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "ae55e9f19a52fdcd97379ac1", "attempt_id": "8d6a3f2e6fbf", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:14:19.710175Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 1132, "message_tokens": 1037, "message_utf8_bytes": 3330, "cache_tensor_bytes": null, "serialized_payload_bytes": 3330, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14577.936624991708, "ttft_ms": 13.601041049696505, "decode_ms": 93.98633299861103, "receiver_prefill_ms": 121.46054196637124}, "total_wall_ms": 14809.023041976616, "ttft_ms": 13.601041049696505, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7782250", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "e756d3601086b09c617d9e1191e356bbef034617cd9082ff9baec27a864a08f2", "output_hash": "35ed445e80874ee97060da399132c68930141ac0b1db960634d510ba5e921356"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "6594814f04aaa26f7ab4f8e6", "attempt_id": "98e54bd6c4eb", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:14:19.959014Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 1125, "message_tokens": 1037, "message_utf8_bytes": 3330, "cache_tensor_bytes": null, "serialized_payload_bytes": 3330, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14577.936624991708, "ttft_ms": 13.328583096154034, "decode_ms": 95.04508308600634, "receiver_prefill_ms": 121.09287502244115}, "total_wall_ms": 14809.431416913867, "ttft_ms": 13.328583096154034, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8846545", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "e756d3601086b09c617d9e1191e356bbef034617cd9082ff9baec27a864a08f2", "output_hash": "f2cac0c994efde1b4c17f7e762ce15722837c5e13b9b535770bb20bde28b8c69"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "9bfdd216305c2d29d7319727", "attempt_id": "3bb91806ca33", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:14:20.208097Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 1134, "message_tokens": 1037, "message_utf8_bytes": 3330, "cache_tensor_bytes": null, "serialized_payload_bytes": 3330, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14577.936624991708, "ttft_ms": 14.434374985285103, "decode_ms": 67.42170790676028, "receiver_prefill_ms": 121.81133299600333}, "total_wall_ms": 14783.606874989346, "ttft_ms": 14.434374985285103, "input_usage": {}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Ruksol-3", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "e756d3601086b09c617d9e1191e356bbef034617cd9082ff9baec27a864a08f2", "output_hash": "df672b6adfdef81742635c0ef2ac39baba09d15dea6c1ec8cdcec73661af11e3"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "551e73a3c90346020940b689", "attempt_id": "df8b0f238973", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:14:20.430778Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 1128, "message_tokens": 1037, "message_utf8_bytes": 3330, "cache_tensor_bytes": null, "serialized_payload_bytes": 3330, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14577.936624991708, "ttft_ms": 13.450167025439441, "decode_ms": 134.28883301094174, "receiver_prefill_ms": 121.18141702376306}, "total_wall_ms": 14848.84700004477, "ttft_ms": 13.450167025439441, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "24,507,488", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "e756d3601086b09c617d9e1191e356bbef034617cd9082ff9baec27a864a08f2", "output_hash": "ace728e4415f8910c352ca013dc105fbafccd2afa834a33285abcbe51a38a305"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "29d093bbc015e55bdd455052", "attempt_id": "6fcf872cea62", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:14:20.718841Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 1128, "message_tokens": 1037, "message_utf8_bytes": 3330, "cache_tensor_bytes": null, "serialized_payload_bytes": 3330, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14577.936624991708, "ttft_ms": 13.89379205647856, "decode_ms": 13.229084084741771, "receiver_prefill_ms": 121.12149992026389}, "total_wall_ms": 14728.254041983746, "ttft_ms": 13.89379205647856, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "e756d3601086b09c617d9e1191e356bbef034617cd9082ff9baec27a864a08f2", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "111a45650e4c887cfbd69e77", "attempt_id": "7d4542469124", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:14:20.886319Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 1130, "message_tokens": 1037, "message_utf8_bytes": 3330, "cache_tensor_bytes": null, "serialized_payload_bytes": 3330, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14577.936624991708, "ttft_ms": 13.665457954630256, "decode_ms": 13.618249911814928, "receiver_prefill_ms": 121.6262080706656}, "total_wall_ms": 14728.813042049296, "ttft_ms": 13.665457954630256, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "e756d3601086b09c617d9e1191e356bbef034617cd9082ff9baec27a864a08f2", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "59b935de6d027cfc2a60a8fd", "attempt_id": "44228ccc0e15", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:14:32.607545Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 10}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 496, "message_tokens": 404, "message_utf8_bytes": 1070, "cache_tensor_bytes": null, "serialized_payload_bytes": 1070, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11552.672000019811, "ttft_ms": 13.478166074492037, "decode_ms": 13.726125005632639, "receiver_prefill_ms": 55.90050003957003}, "total_wall_ms": 11642.658042022958, "ttft_ms": 13.478166074492037, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "26cc7abf857f0365b1fa6a5bf8a4e6a37133624aec654eb7d54c855701b49d6b", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "1510c54290b915634c2cfdbb", "attempt_id": "a8a1213ae04b", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:14:32.715230Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 10}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 497, "message_tokens": 404, "message_utf8_bytes": 1070, "cache_tensor_bytes": null, "serialized_payload_bytes": 1070, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11552.672000019811, "ttft_ms": 13.426749967038631, "decode_ms": 13.453125022351742, "receiver_prefill_ms": 55.7606250513345}, "total_wall_ms": 11636.500875116326, "ttft_ms": 13.426749967038631, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "26cc7abf857f0365b1fa6a5bf8a4e6a37133624aec654eb7d54c855701b49d6b", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "aa0558bcfbbb8e705eede7d7", "attempt_id": "283cf59a2e6e", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:14:32.817319Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 10}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 502, "message_tokens": 404, "message_utf8_bytes": 1070, "cache_tensor_bytes": null, "serialized_payload_bytes": 1070, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11552.672000019811, "ttft_ms": 13.551292009651661, "decode_ms": 13.000417035073042, "receiver_prefill_ms": 56.01170798763633}, "total_wall_ms": 11636.426707962528, "ttft_ms": 13.551292009651661, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "26cc7abf857f0365b1fa6a5bf8a4e6a37133624aec654eb7d54c855701b49d6b", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "75685b36cfd4f2463c7559fd", "attempt_id": "e505e5e23721", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:14:32.918988Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 10}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 495, "message_tokens": 404, "message_utf8_bytes": 1070, "cache_tensor_bytes": null, "serialized_payload_bytes": 1070, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11552.672000019811, "ttft_ms": 13.35283403750509, "decode_ms": 13.296416960656643, "receiver_prefill_ms": 55.82312494516373}, "total_wall_ms": 11636.280999984592, "ttft_ms": 13.35283403750509, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "26cc7abf857f0365b1fa6a5bf8a4e6a37133624aec654eb7d54c855701b49d6b", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b8ad94cc5f70158e524d54fe", "attempt_id": "976d4feb70a6", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:14:33.020235Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 10}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 504, "message_tokens": 404, "message_utf8_bytes": 1070, "cache_tensor_bytes": null, "serialized_payload_bytes": 1070, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11552.672000019811, "ttft_ms": 13.944791979156435, "decode_ms": 79.5171670615673, "receiver_prefill_ms": 55.804750067181885}, "total_wall_ms": 11703.1271669548, "ttft_ms": 13.944791979156435, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Lithzenban-5", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "26cc7abf857f0365b1fa6a5bf8a4e6a37133624aec654eb7d54c855701b49d6b", "output_hash": "cc5e85bb9beafdd57dcb7f11c04a12aea0492c44343bf9633324f1d50e9c4b6f"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "0bae74971fe2445d7fffb9f9", "attempt_id": "6cd97b92852d", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:14:33.188456Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 10}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 498, "message_tokens": 404, "message_utf8_bytes": 1070, "cache_tensor_bytes": null, "serialized_payload_bytes": 1070, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11552.672000019811, "ttft_ms": 13.747040997259319, "decode_ms": 132.7349579660222, "receiver_prefill_ms": 56.08375009614974}, "total_wall_ms": 11756.435542018153, "ttft_ms": 13.747040997259319, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "38,659,135", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "26cc7abf857f0365b1fa6a5bf8a4e6a37133624aec654eb7d54c855701b49d6b", "output_hash": "c49cce3f807e7879e3552dc2ecb0acdb9acb9d80931ed44d38fd9775d45978e6"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "11345299381ebbffea6d7169", "attempt_id": "8c586d07e99f", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:14:33.409815Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 10}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 498, "message_tokens": 404, "message_utf8_bytes": 1070, "cache_tensor_bytes": null, "serialized_payload_bytes": 1070, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11552.672000019811, "ttft_ms": 13.276749989017844, "decode_ms": 13.247375027276576, "receiver_prefill_ms": 55.64637505449355}, "total_wall_ms": 11635.977750062011, "ttft_ms": 13.276749989017844, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "26cc7abf857f0365b1fa6a5bf8a4e6a37133624aec654eb7d54c855701b49d6b", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "ca3ea5b47e9c6d8c94ef52a0", "attempt_id": "ba1b5b6a6ab0", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:14:33.510135Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 10}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 500, "message_tokens": 404, "message_utf8_bytes": 1070, "cache_tensor_bytes": null, "serialized_payload_bytes": 1070, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11552.672000019811, "ttft_ms": 13.329375069588423, "decode_ms": 13.328416971489787, "receiver_prefill_ms": 55.580749991349876}, "total_wall_ms": 11636.037416057661, "ttft_ms": 13.329375069588423, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "26cc7abf857f0365b1fa6a5bf8a4e6a37133624aec654eb7d54c855701b49d6b", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "e50614384f1fc7bb24be00c5", "attempt_id": "e7ffee7a96ce", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:14:45.127190Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 1275, "message_tokens": 1184, "message_utf8_bytes": 3035, "cache_tensor_bytes": null, "serialized_payload_bytes": 3035, "side_information_bytes": 0, "fetched_bytes": 1770, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11516.197458026, "ttft_ms": 13.676707982085645, "decode_ms": 94.62304098997265, "receiver_prefill_ms": 131.78012508433312}, "total_wall_ms": 11758.331624092534, "ttft_ms": 13.676707982085645, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "9519700", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "71314a1a1e3a82c2d9d38865a4ba41c6c58993729e567e593b746373079f3e38", "output_hash": "faba5ef4f83e7fb5277d911a073e9e58b13a2602ac2a2a17da9c4966ab281491"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "42fd79e5790f1ded80b09457", "attempt_id": "2fcd65cdf19b", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:14:45.387565Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 1231, "message_tokens": 1139, "message_utf8_bytes": 2993, "cache_tensor_bytes": null, "serialized_payload_bytes": 2993, "side_information_bytes": 0, "fetched_bytes": 1726, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11516.197458026, "ttft_ms": 13.488416909240186, "decode_ms": 13.377874973230064, "receiver_prefill_ms": 128.1807499472052}, "total_wall_ms": 11673.651125049219, "ttft_ms": 13.488416909240186, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "9feb66525c63e223bb44403b48372d0599013c9c7e87c18cfad693cd96efdb98", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "accefd1fbf14d5de341d4250", "attempt_id": "d6fe448ad600", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:14:45.562458Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 1236, "message_tokens": 1139, "message_utf8_bytes": 2993, "cache_tensor_bytes": null, "serialized_payload_bytes": 2993, "side_information_bytes": 0, "fetched_bytes": 1726, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11516.197458026, "ttft_ms": 13.523667003028095, "decode_ms": 93.60054193530232, "receiver_prefill_ms": 128.416332998313}, "total_wall_ms": 11753.738208091818, "ttft_ms": 13.523667003028095, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7782250", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "9feb66525c63e223bb44403b48372d0599013c9c7e87c18cfad693cd96efdb98", "output_hash": "35ed445e80874ee97060da399132c68930141ac0b1db960634d510ba5e921356"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "289c5aad08ba0e5fc0a1e646", "attempt_id": "5906518dfa4d", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:14:45.817159Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 1229, "message_tokens": 1139, "message_utf8_bytes": 2993, "cache_tensor_bytes": null, "serialized_payload_bytes": 2993, "side_information_bytes": 0, "fetched_bytes": 1726, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11516.197458026, "ttft_ms": 13.321833917871118, "decode_ms": 13.627083040773869, "receiver_prefill_ms": 127.85462499596179}, "total_wall_ms": 11673.05550002493, "ttft_ms": 13.321833917871118, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "9feb66525c63e223bb44403b48372d0599013c9c7e87c18cfad693cd96efdb98", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c57e3a869b90d8df8de2d758", "attempt_id": "391ea5e0ac65", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:14:45.991058Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 1238, "message_tokens": 1139, "message_utf8_bytes": 2993, "cache_tensor_bytes": null, "serialized_payload_bytes": 2993, "side_information_bytes": 0, "fetched_bytes": 1726, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11516.197458026, "ttft_ms": 13.590082991868258, "decode_ms": 80.61983401421458, "receiver_prefill_ms": 128.71816603001207}, "total_wall_ms": 11741.136083030142, "ttft_ms": 13.590082991868258, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Project Lithzenban-5", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a34ac7372702c89371ac5919561e209f54036691348695222a573a7a8f920d8c", "output_hash": "56a237d92e42153f57f3a850b630a2649e422b80d15d714dd5247b068f5f8109"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "66ee79004b5fd5e02966f6d9", "attempt_id": "21e3a57873a9", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:14:46.235714Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 1252, "message_tokens": 1159, "message_utf8_bytes": 3029, "cache_tensor_bytes": null, "serialized_payload_bytes": 3029, "side_information_bytes": 0, "fetched_bytes": 1765, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11516.197458026, "ttft_ms": 13.637166004627943, "decode_ms": 134.28333296906203, "receiver_prefill_ms": 130.9453749563545}, "total_wall_ms": 11797.08104103338, "ttft_ms": 13.637166004627943, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "37,069,563", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "4d6a785f9c614694b251558378026c256051751f1e70dda5129a82fccdf03022", "output_hash": "536437b7f1a51facf14e0322c52f3fcc1f3106369a8c2fe014287a8cbfd4cd64"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "76d4f3dbc02188438e2ec8f9", "attempt_id": "0b470fae6b27", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:14:46.534003Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 1232, "message_tokens": 1139, "message_utf8_bytes": 2993, "cache_tensor_bytes": null, "serialized_payload_bytes": 2993, "side_information_bytes": 0, "fetched_bytes": 1726, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11516.197458026, "ttft_ms": 13.541917083784938, "decode_ms": 13.216375024057925, "receiver_prefill_ms": 128.03741707466543}, "total_wall_ms": 11673.004623968154, "ttft_ms": 13.541917083784938, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "9feb66525c63e223bb44403b48372d0599013c9c7e87c18cfad693cd96efdb98", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c70c079fd962586e72f5288c", "attempt_id": "61695b97a032", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:14:46.708490Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 1234, "message_tokens": 1139, "message_utf8_bytes": 2993, "cache_tensor_bytes": null, "serialized_payload_bytes": 2993, "side_information_bytes": 0, "fetched_bytes": 1726, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11516.197458026, "ttft_ms": 13.43520893715322, "decode_ms": 13.597541954368353, "receiver_prefill_ms": 129.0042910259217}, "total_wall_ms": 11674.303208012134, "ttft_ms": 13.43520893715322, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "9feb66525c63e223bb44403b48372d0599013c9c7e87c18cfad693cd96efdb98", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "294641d240eb10da21e86af2", "attempt_id": "93be17c75553", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:14:47.690168Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 230252544}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 45, "message_tokens": null, "message_utf8_bytes": 30375, "cache_tensor_bytes": 230252544, "serialized_payload_bytes": 230252544, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 806.1215420020744, "ttft_ms": 53.94849996082485, "decode_ms": 138.80883296951652, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1008.0416250275448, "ttft_ms": 53.94849996082485, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "9,519,700", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "f31f51fd5fb1986ddf80fc727e98ae3da80e1335676b76bd1bc1660a441d2eb2"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "907ae3113ea8e5db7564e89e", "attempt_id": "ce10a48a0f1a", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:14:47.909457Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 230252544}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 46, "message_tokens": null, "message_utf8_bytes": 30375, "cache_tensor_bytes": 230252544, "serialized_payload_bytes": 230252544, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 806.1215420020744, "ttft_ms": 46.28258291631937, "decode_ms": 150.34525003284216, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1010.3403760585934, "ttft_ms": 46.28258291631937, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "YA-85D-7885", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "d7c446b684af4f3a0aa8481919c405aba8a334272947b087f62b1b6c80b3f946"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "9e53bb66ec58251a35d89442", "attempt_id": "b8659dba05a7", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:14:48.131194Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 230252544}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 51, "message_tokens": null, "message_utf8_bytes": 30375, "cache_tensor_bytes": 230252544, "serialized_payload_bytes": 230252544, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 806.1215420020744, "ttft_ms": 48.722082981839776, "decode_ms": 104.53329200390726, "receiver_prefill_ms": 0.0}, "total_wall_ms": 965.6549590872601, "ttft_ms": 48.722082981839776, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7782250", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "35ed445e80874ee97060da399132c68930141ac0b1db960634d510ba5e921356"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "246a4c7798f95b422c952762", "attempt_id": "deb311190814", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:14:48.308052Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 230252544}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 44, "message_tokens": null, "message_utf8_bytes": 30375, "cache_tensor_bytes": 230252544, "serialized_payload_bytes": 230252544, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 806.1215420020744, "ttft_ms": 47.16312501113862, "decode_ms": 120.24020904209465, "receiver_prefill_ms": 0.0}, "total_wall_ms": 980.8885839302093, "ttft_ms": 47.16312501113862, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "10897865", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "795facd6ec9bf870c1fd5245bfde895e2cdf0e6690ae28308b78883d80e86927"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "2f464af296078c59e4cecf58", "attempt_id": "034b8517e2ce", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:14:48.500212Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 230252544}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 53, "message_tokens": null, "message_utf8_bytes": 30375, "cache_tensor_bytes": 230252544, "serialized_payload_bytes": 230252544, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 806.1215420020744, "ttft_ms": 48.66233398206532, "decode_ms": 92.29608392342925, "receiver_prefill_ms": 0.0}, "total_wall_ms": 954.492500051856, "ttft_ms": 48.66233398206532, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Lithzenban-5", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "cc5e85bb9beafdd57dcb7f11c04a12aea0492c44343bf9633324f1d50e9c4b6f"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "140aec4309ffe46bf50140c3", "attempt_id": "b340e2d29e62", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:14:48.665989Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 230252544}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 30375, "cache_tensor_bytes": 230252544, "serialized_payload_bytes": 230252544, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 806.1215420020744, "ttft_ms": 47.15662507805973, "decode_ms": 120.07404095493257, "receiver_prefill_ms": 0.0}, "total_wall_ms": 980.6722090579569, "ttft_ms": 47.15662507805973, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "23747418", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "7bd2d5f426d52effec1f9f18ad1abf8f6f1f3d45bed737f2493898f8f290fa69"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "61b7c08d9e600a023939ed68", "attempt_id": "2b37a240480c", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:14:48.858329Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 230252544}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 30375, "cache_tensor_bytes": 230252544, "serialized_payload_bytes": 230252544, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 806.1215420020744, "ttft_ms": 47.11566597688943, "decode_ms": 15.030000009573996, "receiver_prefill_ms": 0.0}, "total_wall_ms": 875.6118749734014, "ttft_ms": 47.11566597688943, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "67af4645835a320c063d1bf9", "attempt_id": "2e9c19560c4d", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:14:48.944802Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 230252544}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 49, "message_tokens": null, "message_utf8_bytes": 30375, "cache_tensor_bytes": 230252544, "serialized_payload_bytes": 230252544, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 806.1215420020744, "ttft_ms": 48.4706669813022, "decode_ms": 15.281333937309682, "receiver_prefill_ms": 0.0}, "total_wall_ms": 877.19308398664, "ttft_ms": 48.4706669813022, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "41e295e0f06ee98ba7114585", "attempt_id": "84793e086299", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:14:49.033124Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 6369, "message_tokens": 6282, "message_utf8_bytes": 30905, "cache_tensor_bytes": null, "serialized_payload_bytes": 30905, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.028875074349343777, "ttft_ms": 14.216125011444092, "decode_ms": 96.73474996816367, "receiver_prefill_ms": 757.3737499769777}, "total_wall_ms": 878.5683331079781, "ttft_ms": 14.216125011444092, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "9519700", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "cbab481407f3b200fbd59486c609f33654dce0feb421e39a933638173957c614", "output_hash": "faba5ef4f83e7fb5277d911a073e9e58b13a2602ac2a2a17da9c4966ab281491"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "77056bf2c6bbc47325e46b0c", "attempt_id": "26720f8463c1", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:14:49.928039Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 6370, "message_tokens": 6282, "message_utf8_bytes": 30905, "cache_tensor_bytes": null, "serialized_payload_bytes": 30905, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.028875074349343777, "ttft_ms": 13.893292052671313, "decode_ms": 138.4029999608174, "receiver_prefill_ms": 778.3742080209777}, "total_wall_ms": 943.1697081308812, "ttft_ms": 13.893292052671313, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "YA-85D-7885", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "cbab481407f3b200fbd59486c609f33654dce0feb421e39a933638173957c614", "output_hash": "d7c446b684af4f3a0aa8481919c405aba8a334272947b087f62b1b6c80b3f946"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "536d9c0d3c76505ccdf64fc1", "attempt_id": "13448c484dea", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:14:50.887761Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 6375, "message_tokens": 6282, "message_utf8_bytes": 30905, "cache_tensor_bytes": null, "serialized_payload_bytes": 30905, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.028875074349343777, "ttft_ms": 13.980000047013164, "decode_ms": 97.72029204759747, "receiver_prefill_ms": 786.6485000122339}, "total_wall_ms": 908.1427081255242, "ttft_ms": 13.980000047013164, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7782250", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "cbab481407f3b200fbd59486c609f33654dce0feb421e39a933638173957c614", "output_hash": "35ed445e80874ee97060da399132c68930141ac0b1db960634d510ba5e921356"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b5d63b3ddfbb37c2e9b48aca", "attempt_id": "d7e4a30a88ea", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:14:51.815504Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 6368, "message_tokens": 6282, "message_utf8_bytes": 30905, "cache_tensor_bytes": null, "serialized_payload_bytes": 30905, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.028875074349343777, "ttft_ms": 14.20324994251132, "decode_ms": 125.34587504342198, "receiver_prefill_ms": 822.248708922416}, "total_wall_ms": 971.6051671421155, "ttft_ms": 14.20324994251132, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8,846,545", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "cbab481407f3b200fbd59486c609f33654dce0feb421e39a933638173957c614", "output_hash": "77d7beaf301ce085d0b52f1ffccc70fe340d883e4f7677ef3ec61ef496cd7d5d"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "f090a532558874c8b95c4727", "attempt_id": "09e6b8106e06", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:14:52.804479Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 6377, "message_tokens": 6282, "message_utf8_bytes": 30905, "cache_tensor_bytes": null, "serialized_payload_bytes": 30905, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.028875074349343777, "ttft_ms": 14.521708013489842, "decode_ms": 83.6466250475496, "receiver_prefill_ms": 800.1294999849051}, "total_wall_ms": 908.1460001179948, "ttft_ms": 14.521708013489842, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Lithzenban-5", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "cbab481407f3b200fbd59486c609f33654dce0feb421e39a933638173957c614", "output_hash": "cc5e85bb9beafdd57dcb7f11c04a12aea0492c44343bf9633324f1d50e9c4b6f"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "19a15882e4dab0e5bc75ad1a", "attempt_id": "be6c5c3e82ba", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:14:53.730109Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 6371, "message_tokens": 6282, "message_utf8_bytes": 30905, "cache_tensor_bytes": null, "serialized_payload_bytes": 30905, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.028875074349343777, "ttft_ms": 14.051332953386009, "decode_ms": 99.0746250608936, "receiver_prefill_ms": 871.2381669320166}, "total_wall_ms": 994.094708119519, "ttft_ms": 14.051332953386009, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4741451", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "cbab481407f3b200fbd59486c609f33654dce0feb421e39a933638173957c614", "output_hash": "f7f74f0b7f27a3da2bacb12f5d772280d8633ae8b202af406bbaede06fd21819"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a162241dfe646a64090fb113", "attempt_id": "4f4c5032d7d0", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:14:54.741505Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 6371, "message_tokens": 6282, "message_utf8_bytes": 30905, "cache_tensor_bytes": null, "serialized_payload_bytes": 30905, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.028875074349343777, "ttft_ms": 14.016375062055886, "decode_ms": 14.017250039614737, "receiver_prefill_ms": 803.3125000074506}, "total_wall_ms": 841.1578751401976, "ttft_ms": 14.016375062055886, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "cbab481407f3b200fbd59486c609f33654dce0feb421e39a933638173957c614", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "382f3cf4770dd9f7280494bb", "attempt_id": "252d45453005", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:14:55.599729Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 6373, "message_tokens": 6282, "message_utf8_bytes": 30905, "cache_tensor_bytes": null, "serialized_payload_bytes": 30905, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.028875074349343777, "ttft_ms": 14.381999964825809, "decode_ms": 13.917083968408406, "receiver_prefill_ms": 825.3269579727203}, "total_wall_ms": 863.5366660309955, "ttft_ms": 14.381999964825809, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "cbab481407f3b200fbd59486c609f33654dce0feb421e39a933638173957c614", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "5d754e49c284f80f2359e5d7", "attempt_id": "6a2cc6b942ac", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:15:23.070705Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1914}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 2028, "message_tokens": 1940, "message_utf8_bytes": 6293, "cache_tensor_bytes": null, "serialized_payload_bytes": 6293, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26589.627083041705, "ttft_ms": 13.539666077122092, "decode_ms": 121.14820897113532, "receiver_prefill_ms": 211.84891695156693}, "total_wall_ms": 26940.360792097636, "ttft_ms": 13.539666077122092, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "9,519,700", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "627a266ed521302dd0777e0144f13268fa4f8f1232c230f7956ddd26712d2a45", "output_hash": "f31f51fd5fb1986ddf80fc727e98ae3da80e1335676b76bd1bc1660a441d2eb2"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "3ad61f16059c241c4037eb2b", "attempt_id": "f456140546fe", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:15:23.439960Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1914}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 2029, "message_tokens": 1940, "message_utf8_bytes": 6293, "cache_tensor_bytes": null, "serialized_payload_bytes": 6293, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26589.627083041705, "ttft_ms": 13.719749986194074, "decode_ms": 135.43158292304724, "receiver_prefill_ms": 211.4697500364855}, "total_wall_ms": 26953.36645806674, "ttft_ms": 13.719749986194074, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "YA-85D-7885", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "627a266ed521302dd0777e0144f13268fa4f8f1232c230f7956ddd26712d2a45", "output_hash": "d7c446b684af4f3a0aa8481919c405aba8a334272947b087f62b1b6c80b3f946"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "3c7cab4dd6c1b2d5947eab90", "attempt_id": "468d61848429", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:15:23.821387Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1914}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 2034, "message_tokens": 1940, "message_utf8_bytes": 6293, "cache_tensor_bytes": null, "serialized_payload_bytes": 6293, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26589.627083041705, "ttft_ms": 13.523499947041273, "decode_ms": 94.0131249371916, "receiver_prefill_ms": 212.9213329171762}, "total_wall_ms": 26913.17583306227, "ttft_ms": 13.523499947041273, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7782250", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "627a266ed521302dd0777e0144f13268fa4f8f1232c230f7956ddd26712d2a45", "output_hash": "35ed445e80874ee97060da399132c68930141ac0b1db960634d510ba5e921356"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "4cf4caf038ba6e8913e23ab4", "attempt_id": "f422a5ee794e", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:15:24.161940Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1914}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 2027, "message_tokens": 1940, "message_utf8_bytes": 6293, "cache_tensor_bytes": null, "serialized_payload_bytes": 6293, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26589.627083041705, "ttft_ms": 13.576167053543031, "decode_ms": 135.42541698552668, "receiver_prefill_ms": 215.55654203984886}, "total_wall_ms": 26957.211750093848, "ttft_ms": 13.576167053543031, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "10,897,865", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "627a266ed521302dd0777e0144f13268fa4f8f1232c230f7956ddd26712d2a45", "output_hash": "bf83aaa75e325e289601f3e10f3355c6cb797de4ffc4ee604831642a5b3ee46b"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "38a69aedbf29b27ded2a1bce", "attempt_id": "5925fec80f2e", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:15:24.547975Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1914}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 2036, "message_tokens": 1940, "message_utf8_bytes": 6293, "cache_tensor_bytes": null, "serialized_payload_bytes": 6293, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26589.627083041705, "ttft_ms": 14.211875037290156, "decode_ms": 82.17116596642882, "receiver_prefill_ms": 232.4280000757426}, "total_wall_ms": 26921.43120802939, "ttft_ms": 14.211875037290156, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Project Lithzenban-5", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "627a266ed521302dd0777e0144f13268fa4f8f1232c230f7956ddd26712d2a45", "output_hash": "56a237d92e42153f57f3a850b630a2649e422b80d15d714dd5247b068f5f8109"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "788fd39343675c5e1fb59bb0", "attempt_id": "05d050492610", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:15:24.897860Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1914}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 2030, "message_tokens": 1940, "message_utf8_bytes": 6293, "cache_tensor_bytes": null, "serialized_payload_bytes": 6293, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26589.627083041705, "ttft_ms": 13.745374977588654, "decode_ms": 135.50637499429286, "receiver_prefill_ms": 231.97850002907217}, "total_wall_ms": 26973.9547080826, "ttft_ms": 13.745374977588654, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "27,293,141", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "627a266ed521302dd0777e0144f13268fa4f8f1232c230f7956ddd26712d2a45", "output_hash": "65c5a3119e648db0be3998eb669de03e43858975d821c49e963df79f9fdc104d"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "27e1bc1ea5cd2bb669a17bea", "attempt_id": "b94b1ba4b424", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:15:25.300439Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1914}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 2030, "message_tokens": 1940, "message_utf8_bytes": 6293, "cache_tensor_bytes": null, "serialized_payload_bytes": 6293, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26589.627083041705, "ttft_ms": 13.74645798932761, "decode_ms": 13.569249887950718, "receiver_prefill_ms": 221.66412498336285}, "total_wall_ms": 26841.641541104764, "ttft_ms": 13.74645798932761, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "627a266ed521302dd0777e0144f13268fa4f8f1232c230f7956ddd26712d2a45", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "0422326c6c94c60e1512f0b3", "attempt_id": "2bbfaa40b961", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:15:25.570246Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1914}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 2032, "message_tokens": 1940, "message_utf8_bytes": 6293, "cache_tensor_bytes": null, "serialized_payload_bytes": 6293, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26589.627083041705, "ttft_ms": 13.662416022270918, "decode_ms": 13.589999987743795, "receiver_prefill_ms": 234.30150002241135}, "total_wall_ms": 26854.198624030687, "ttft_ms": 13.662416022270918, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "627a266ed521302dd0777e0144f13268fa4f8f1232c230f7956ddd26712d2a45", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "d2bdbba3f5c99da1202bc227", "attempt_id": "c7ab7ae7c5b1", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:15:48.316491Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 22}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 949, "message_tokens": 857, "message_utf8_bytes": 2286, "cache_tensor_bytes": null, "serialized_payload_bytes": 2286, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 22464.10095889587, "ttft_ms": 13.522083987481892, "decode_ms": 13.321708072908223, "receiver_prefill_ms": 391.7809169506654}, "total_wall_ms": 22884.43741784431, "ttft_ms": 13.522083987481892, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "3ab26c6f38988d38e0dda0c3d182a3dccd83915dce526ffee0d6688047c14a94", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "feff3ecf229b36ed010fc51f", "attempt_id": "b86f3789d70d", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:15:48.756400Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 22}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 950, "message_tokens": 857, "message_utf8_bytes": 2286, "cache_tensor_bytes": null, "serialized_payload_bytes": 2286, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 22464.10095889587, "ttft_ms": 21.749084000475705, "decode_ms": 13.953874935396016, "receiver_prefill_ms": 162.1962080243975}, "total_wall_ms": 22663.699750904925, "ttft_ms": 21.749084000475705, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "3ab26c6f38988d38e0dda0c3d182a3dccd83915dce526ffee0d6688047c14a94", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a4890aceb850b4269cf1ea51", "attempt_id": "e283b13a794b", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:15:48.974365Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 22}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 955, "message_tokens": 857, "message_utf8_bytes": 2286, "cache_tensor_bytes": null, "serialized_payload_bytes": 2286, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 22464.10095889587, "ttft_ms": 14.097249950282276, "decode_ms": 13.740292051807046, "receiver_prefill_ms": 112.05062502995133}, "total_wall_ms": 22605.99058389198, "ttft_ms": 14.097249950282276, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "3ab26c6f38988d38e0dda0c3d182a3dccd83915dce526ffee0d6688047c14a94", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "4d47d08189494f1c58df9bcf", "attempt_id": "075b7561b04f", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:15:49.135010Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 22}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 948, "message_tokens": 857, "message_utf8_bytes": 2286, "cache_tensor_bytes": null, "serialized_payload_bytes": 2286, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 22464.10095889587, "ttft_ms": 14.386166003532708, "decode_ms": 98.048708983697, "receiver_prefill_ms": 103.74270798638463}, "total_wall_ms": 22684.59058389999, "ttft_ms": 14.386166003532708, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8846545", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "3ab26c6f38988d38e0dda0c3d182a3dccd83915dce526ffee0d6688047c14a94", "output_hash": "f2cac0c994efde1b4c17f7e762ce15722837c5e13b9b535770bb20bde28b8c69"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "7484c7c26f34ebf15b5ce113", "attempt_id": "e3db51f7ab16", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:15:49.374587Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 22}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 957, "message_tokens": 857, "message_utf8_bytes": 2286, "cache_tensor_bytes": null, "serialized_payload_bytes": 2286, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 22464.10095889587, "ttft_ms": 14.011749997735023, "decode_ms": 71.48408307693899, "receiver_prefill_ms": 104.12608401384205}, "total_wall_ms": 22656.92091686651, "ttft_ms": 14.011749997735023, "input_usage": {}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Ruksol-3", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "3ab26c6f38988d38e0dda0c3d182a3dccd83915dce526ffee0d6688047c14a94", "output_hash": "df672b6adfdef81742635c0ef2ac39baba09d15dea6c1ec8cdcec73661af11e3"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "af1c8c6c9c22aa75224b0be0", "attempt_id": "e1a89154726e", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:15:49.587772Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 22}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 951, "message_tokens": 857, "message_utf8_bytes": 2286, "cache_tensor_bytes": null, "serialized_payload_bytes": 2286, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 22464.10095889587, "ttft_ms": 16.308833030052483, "decode_ms": 111.568458029069, "receiver_prefill_ms": 107.34099999535829}, "total_wall_ms": 22701.721500838175, "ttft_ms": 16.308833030052483, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1878125", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "3ab26c6f38988d38e0dda0c3d182a3dccd83915dce526ffee0d6688047c14a94", "output_hash": "7c02acfbb5cc2bb56f6a72e9c70ecd67c916a2cce64b5a518c3437773bd9f819"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "638264574b3a53ea525eebdf", "attempt_id": "16defd0772e2", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:15:49.850841Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 22}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 951, "message_tokens": 857, "message_utf8_bytes": 2286, "cache_tensor_bytes": null, "serialized_payload_bytes": 2286, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 22464.10095889587, "ttft_ms": 14.30229190737009, "decode_ms": 14.75649990607053, "receiver_prefill_ms": 111.59591597970575}, "total_wall_ms": 22606.779917958193, "ttft_ms": 14.30229190737009, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "3ab26c6f38988d38e0dda0c3d182a3dccd83915dce526ffee0d6688047c14a94", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "0087ed7765a13ece62a84200", "attempt_id": "44ba04a6db22", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:15:50.012354Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 22}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 953, "message_tokens": 857, "message_utf8_bytes": 2286, "cache_tensor_bytes": null, "serialized_payload_bytes": 2286, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 22464.10095889587, "ttft_ms": 13.90516699757427, "decode_ms": 13.527749921195209, "receiver_prefill_ms": 116.65541701950133}, "total_wall_ms": 22610.114917857572, "ttft_ms": 13.90516699757427, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "3ab26c6f38988d38e0dda0c3d182a3dccd83915dce526ffee0d6688047c14a94", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "31716e6077699603747f2919", "attempt_id": "5f183325eca7", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:16:12.145284Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 2451, "message_tokens": 2360, "message_utf8_bytes": 6102, "cache_tensor_bytes": null, "serialized_payload_bytes": 6102, "side_information_bytes": 0, "fetched_bytes": 3450, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 21968.349708011374, "ttft_ms": 13.496499974280596, "decode_ms": 94.66599998995662, "receiver_prefill_ms": 355.49112502485514}, "total_wall_ms": 22435.923250042833, "ttft_ms": 13.496499974280596, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "9519700", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "f268c83a76e13afe81e3fb36d51d58a3d141c8b77fe362f4a5ce7a8e04c980c5", "output_hash": "faba5ef4f83e7fb5277d911a073e9e58b13a2602ac2a2a17da9c4966ab281491"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "28b839ce8c8c7ad2e2200476", "attempt_id": "8f272417c9a2", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:16:12.639472Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 2421, "message_tokens": 2329, "message_utf8_bytes": 6048, "cache_tensor_bytes": null, "serialized_payload_bytes": 6048, "side_information_bytes": 0, "fetched_bytes": 3407, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 21968.349708011374, "ttft_ms": 13.47799994982779, "decode_ms": 135.34929207526147, "receiver_prefill_ms": 270.6201249966398}, "total_wall_ms": 22391.45058300346, "ttft_ms": 13.47799994982779, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "FY-37T-8239", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "13c31f2b4ff63d6588b67956b8f2dcf3df30fe01ef6b88c11cea34cf68ea24a2", "output_hash": "ec854d998232bb8dc5303bad96e3e9c600d7e9fbf301973d1a9aefbd7b23dbf9"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "7ce6d370f7dc7b814a18a6ce", "attempt_id": "50b851a0907d", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:16:13.080010Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 2457, "message_tokens": 2360, "message_utf8_bytes": 6102, "cache_tensor_bytes": null, "serialized_payload_bytes": 6102, "side_information_bytes": 0, "fetched_bytes": 3450, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 21968.349708011374, "ttft_ms": 14.138167025521398, "decode_ms": 13.955791015177965, "receiver_prefill_ms": 317.7531249821186}, "total_wall_ms": 22317.934542079456, "ttft_ms": 14.138167025521398, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "e8ad56be127551942881f321118074861dde3b8989c6c39cdc6a8dbe25a6b6cc", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "55e35dc5e83ab787fb56d0b1", "attempt_id": "8be83cc11040", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:16:13.448176Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 2451, "message_tokens": 2361, "message_utf8_bytes": 6103, "cache_tensor_bytes": null, "serialized_payload_bytes": 6103, "side_information_bytes": 0, "fetched_bytes": 3450, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 21968.349708011374, "ttft_ms": 13.82195798214525, "decode_ms": 95.61337495688349, "receiver_prefill_ms": 307.06845794338733}, "total_wall_ms": 22388.348083011806, "ttft_ms": 13.82195798214525, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8846545", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "15e925d10af5989f4605d7e5defb3334ef15ab6dbea85738a98a0dbfce116be8", "output_hash": "f2cac0c994efde1b4c17f7e762ce15722837c5e13b9b535770bb20bde28b8c69"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "ef275aa1d161016a04026aee", "attempt_id": "1ac1c08fa799", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:16:13.886720Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 2415, "message_tokens": 2316, "message_utf8_bytes": 6041, "cache_tensor_bytes": null, "serialized_payload_bytes": 6041, "side_information_bytes": 0, "fetched_bytes": 3379, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 21968.349708011374, "ttft_ms": 14.059875044040382, "decode_ms": 69.0913749858737, "receiver_prefill_ms": 292.99541702494025}, "total_wall_ms": 22347.898249980062, "ttft_ms": 14.059875044040382, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Ruksol-3", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "db210b2f7ee6de7ba3dd8513e9e93c1e31421248e596e8854196f82b067076cb", "output_hash": "df672b6adfdef81742635c0ef2ac39baba09d15dea6c1ec8cdcec73661af11e3"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c5a48813db63830a86661ba2", "attempt_id": "a37848606c88", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:16:14.284224Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 2453, "message_tokens": 2360, "message_utf8_bytes": 6102, "cache_tensor_bytes": null, "serialized_payload_bytes": 6102, "side_information_bytes": 0, "fetched_bytes": 3450, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 21968.349708011374, "ttft_ms": 13.780917041003704, "decode_ms": 139.86449991352856, "receiver_prefill_ms": 294.140666956082}, "total_wall_ms": 22419.653665972874, "ttft_ms": 13.780917041003704, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "27,781,038", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "f268c83a76e13afe81e3fb36d51d58a3d141c8b77fe362f4a5ce7a8e04c980c5", "output_hash": "5703ce3ad678d538852d0c714145dae8b5c8481460970cb9c26befe7c98ecb07"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "f78035e6cb5bea99e0c3ae73", "attempt_id": "07055baa41da", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:16:14.753909Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 2440, "message_tokens": 2347, "message_utf8_bytes": 6084, "cache_tensor_bytes": null, "serialized_payload_bytes": 6084, "side_information_bytes": 0, "fetched_bytes": 3430, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 21968.349708011374, "ttft_ms": 14.362709014676511, "decode_ms": 14.586749952286482, "receiver_prefill_ms": 363.5609591146931}, "total_wall_ms": 22364.302542060614, "ttft_ms": 14.362709014676511, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "72ba8932f9b9e8abe99e7ce986f5e940c25f74d6dd43f32b6a1a5c34e60bb02a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "694fc00bd310d4a0ca9d5623", "attempt_id": "f81f47b758dc", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:16:15.168490Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 2455, "message_tokens": 2360, "message_utf8_bytes": 6102, "cache_tensor_bytes": null, "serialized_payload_bytes": 6102, "side_information_bytes": 0, "fetched_bytes": 3450, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 21968.349708011374, "ttft_ms": 13.886332977563143, "decode_ms": 13.637542026117444, "receiver_prefill_ms": 318.65300005301833}, "total_wall_ms": 22317.98204104416, "ttft_ms": 13.886332977563143, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "e8ad56be127551942881f321118074861dde3b8989c6c39cdc6a8dbe25a6b6cc", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "4387663fe5d8704ca6089ef4", "attempt_id": "7ea4fc2712e3", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:16:16.456183Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 237330432}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 45, "message_tokens": null, "message_utf8_bytes": 30851, "cache_tensor_bytes": 237330432, "serialized_payload_bytes": 237330432, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 919.4385420996696, "ttft_ms": 91.36808302719146, "decode_ms": 107.60250000748783, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1132.1084591327235, "ttft_ms": 91.36808302719146, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "9519700", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "faba5ef4f83e7fb5277d911a073e9e58b13a2602ac2a2a17da9c4966ab281491"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "4d17d062976860f86ae75c16", "attempt_id": "98042c6ee1bf", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:16:16.687048Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 237330432}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 46, "message_tokens": null, "message_utf8_bytes": 30851, "cache_tensor_bytes": 237330432, "serialized_payload_bytes": 237330432, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 919.4385420996696, "ttft_ms": 83.82825006265193, "decode_ms": 151.34841692633927, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1159.923208062537, "ttft_ms": 83.82825006265193, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "YA-85D-7885", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "d7c446b684af4f3a0aa8481919c405aba8a334272947b087f62b1b6c80b3f946"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "d489f1491702697b2ba71e63", "attempt_id": "cb7dc0938ffd", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:16:16.945694Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 237330432}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 51, "message_tokens": null, "message_utf8_bytes": 30851, "cache_tensor_bytes": 237330432, "serialized_payload_bytes": 237330432, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 919.4385420996696, "ttft_ms": 136.01729203946888, "decode_ms": 144.49158299248666, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1204.759709071368, "ttft_ms": 136.01729203946888, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "9,657,480", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "c662895167f7326c65e57f72b542b1352330eb99d896451e1796a06d8ea72c18"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "165186e4d0a0d05cac27f1e0", "attempt_id": "433983ab96c9", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:16:17.248808Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 237330432}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 44, "message_tokens": null, "message_utf8_bytes": 30851, "cache_tensor_bytes": 237330432, "serialized_payload_bytes": 237330432, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 919.4385420996696, "ttft_ms": 57.13187495712191, "decode_ms": 151.37012500781566, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1132.842417107895, "ttft_ms": 57.13187495712191, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "10,511,705", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "fd5c4c1519a88ceaebebe6d5443ebb2122f0b6a44bd1baca9b0ad92384eb0dc3"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b790ffb5a121cf987cae8045", "attempt_id": "28e322f3d0a6", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:16:17.480112Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 237330432}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 53, "message_tokens": null, "message_utf8_bytes": 30851, "cache_tensor_bytes": 237330432, "serialized_payload_bytes": 237330432, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 919.4385420996696, "ttft_ms": 50.303875003010035, "decode_ms": 91.10712504480034, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1066.10191706568, "ttft_ms": 50.303875003010035, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Lithzenban-5", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "cc5e85bb9beafdd57dcb7f11c04a12aea0492c44343bf9633324f1d50e9c4b6f"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "f0c78919877e96dd59483de0", "attempt_id": "80a93f9bf7b7", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:16:17.644778Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 237330432}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 30851, "cache_tensor_bytes": 237330432, "serialized_payload_bytes": 237330432, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 919.4385420996696, "ttft_ms": 48.68891602382064, "decode_ms": 121.07395904604346, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1094.558000098914, "ttft_ms": 48.68891602382064, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "18623415", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "c2690dd392cee29b6d09719fc3f49970fbfb1172af27cd49ac3142b269d4f203"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "1387a418020a0d7f9f7805a9", "attempt_id": "a8dc55027b95", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:16:17.838204Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 237330432}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 30851, "cache_tensor_bytes": 237330432, "serialized_payload_bytes": 237330432, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 919.4385420996696, "ttft_ms": 48.39404101949185, "decode_ms": 15.330041060224175, "receiver_prefill_ms": 0.0}, "total_wall_ms": 988.4263331769034, "ttft_ms": 48.39404101949185, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "126a88f02c7740b1ed4814b3", "attempt_id": "31735f6f63e8", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:16:17.925663Z", "task_family": "synthetic", "document_id": "syn_03013_8000", "document_hash": "779b3b1693ab26a564a0bbea9239f13134bc3d34d4fb5688b86da554a5994b05", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03013_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 237330432}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5810, "query_tokens": 49, "message_tokens": null, "message_utf8_bytes": 30851, "cache_tensor_bytes": 237330432, "serialized_payload_bytes": 237330432, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 919.4385420996696, "ttft_ms": 49.74783398211002, "decode_ms": 15.084125101566315, "receiver_prefill_ms": 0.0}, "total_wall_ms": 989.7640420822427, "ttft_ms": 49.74783398211002, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lithzenban-5", "bogus_amount": "8,422,765", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03013_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "87e793faba2396832a7859bc", "attempt_id": "4fd3e4e94288", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:16:18.977530Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 2306, "message_tokens": 2219, "message_utf8_bytes": 8104, "cache_tensor_bytes": null, "serialized_payload_bytes": 8104, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.030499999411404133, "ttft_ms": 13.610666966997087, "decode_ms": 94.50762497726828, "receiver_prefill_ms": 269.2921250127256}, "total_wall_ms": 382.037749979645, "ttft_ms": 13.610666966997087, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8421725", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "46521933ce5dc46ca81ac612352e4fd643797d4a747c5d277c4d2fd41a973767", "output_hash": "e4949aba3b0b77de05a87edaab854329fa75df7bf68b29a3def71e985a06e5b5"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "3a34cebd33d7b227443284ed", "attempt_id": "9ffa0aa180c3", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:16:19.378633Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 2306, "message_tokens": 2219, "message_utf8_bytes": 8104, "cache_tensor_bytes": null, "serialized_payload_bytes": 8104, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.030499999411404133, "ttft_ms": 13.937584008090198, "decode_ms": 136.10870798584074, "receiver_prefill_ms": 257.24687497131526}, "total_wall_ms": 410.9183329856023, "ttft_ms": 13.937584008090198, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "ZW-73S-2197", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "46521933ce5dc46ca81ac612352e4fd643797d4a747c5d277c4d2fd41a973767", "output_hash": "5ab36a19a647e8707dc0e01150dbfc0b49554c16496650f6b32f847c5e5bdf04"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "13ecf177f588e38ad16126bd", "attempt_id": "5efdad76474d", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:16:19.803762Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 2312, "message_tokens": 2219, "message_utf8_bytes": 8104, "cache_tensor_bytes": null, "serialized_payload_bytes": 8104, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.030499999411404133, "ttft_ms": 13.547042035497725, "decode_ms": 121.39012501575053, "receiver_prefill_ms": 257.5641250004992}, "total_wall_ms": 396.2383749894798, "ttft_ms": 13.547042035497725, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "3,199,125", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "46521933ce5dc46ca81ac612352e4fd643797d4a747c5d277c4d2fd41a973767", "output_hash": "f5f7ba5ec2f9b2364ec9e09349c810ade23d11404d04c3b17cea227797de4e79"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c35d3af36ee9ed8a43e61027", "attempt_id": "276c529f707a", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:16:20.214593Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 2306, "message_tokens": 2219, "message_utf8_bytes": 8104, "cache_tensor_bytes": null, "serialized_payload_bytes": 8104, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.030499999411404133, "ttft_ms": 13.87187500949949, "decode_ms": 121.9810419715941, "receiver_prefill_ms": 256.97020906955004}, "total_wall_ms": 396.49820898193866, "ttft_ms": 13.87187500949949, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7,791,055", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "46521933ce5dc46ca81ac612352e4fd643797d4a747c5d277c4d2fd41a973767", "output_hash": "a7a9e7d4ae8669152180c3b8a16087d50f29ecb3daa859b35b9aef6996508884"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "8489e13c9c964b0f0ba8a92b", "attempt_id": "c46e03d6bbe1", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:16:20.625667Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 2315, "message_tokens": 2219, "message_utf8_bytes": 8104, "cache_tensor_bytes": null, "serialized_payload_bytes": 8104, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.030499999411404133, "ttft_ms": 15.12908295262605, "decode_ms": 82.30216708034277, "receiver_prefill_ms": 258.4344580536708}, "total_wall_ms": 359.65425008907914, "ttft_ms": 15.12908295262605, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Pragamfel-2", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "46521933ce5dc46ca81ac612352e4fd643797d4a747c5d277c4d2fd41a973767", "output_hash": "f0df32e3152d54160c1f3bdeb20ad87cbc7e1efcd3ea28c84ac1d2be0ed97c7a"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b9656a935a6bedcca97d422e", "attempt_id": "cb2f828cfb3b", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:16:20.999113Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 2308, "message_tokens": 2219, "message_utf8_bytes": 8104, "cache_tensor_bytes": null, "serialized_payload_bytes": 8104, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.030499999411404133, "ttft_ms": 13.73916701413691, "decode_ms": 95.24604200851172, "receiver_prefill_ms": 256.7931659286842}, "total_wall_ms": 369.492374942638, "ttft_ms": 13.73916701413691, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1144450", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "46521933ce5dc46ca81ac612352e4fd643797d4a747c5d277c4d2fd41a973767", "output_hash": "ecba73db17420cb4334e1cdf134583c639a0ba2bd6f41436035ff2b05291629d"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "2013b11f148ebfadaced1ac1", "attempt_id": "8d0b86e84942", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:16:21.383136Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 2309, "message_tokens": 2219, "message_utf8_bytes": 8104, "cache_tensor_bytes": null, "serialized_payload_bytes": 8104, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.030499999411404133, "ttft_ms": 13.368209009058774, "decode_ms": 13.599542086012661, "receiver_prefill_ms": 257.2214170359075}, "total_wall_ms": 287.7745000878349, "ttft_ms": 13.368209009058774, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "46521933ce5dc46ca81ac612352e4fd643797d4a747c5d277c4d2fd41a973767", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "50afa026dd62cc3b6fc34894", "attempt_id": "3d888e81d343", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:16:21.685336Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 2310, "message_tokens": 2219, "message_utf8_bytes": 8104, "cache_tensor_bytes": null, "serialized_payload_bytes": 8104, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.030499999411404133, "ttft_ms": 13.515708036720753, "decode_ms": 13.83095805067569, "receiver_prefill_ms": 257.5491670286283}, "total_wall_ms": 288.6473749531433, "ttft_ms": 13.515708036720753, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "46521933ce5dc46ca81ac612352e4fd643797d4a747c5d277c4d2fd41a973767", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "f6670ef3b47a25e72232a421", "attempt_id": "34de0ea0e501", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:16:36.044839Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 1124, "message_tokens": 1036, "message_utf8_bytes": 2522, "cache_tensor_bytes": null, "serialized_payload_bytes": 2522, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14056.77837494295, "ttft_ms": 13.329792069271207, "decode_ms": 93.60316605307162, "receiver_prefill_ms": 121.64616701193154}, "total_wall_ms": 14292.438916978426, "ttft_ms": 13.329792069271207, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8421725", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "91aab8fec01151a17508e3b9fa4bc9164d05f5ba8bf8bfc8c47f78922bba670a", "output_hash": "e4949aba3b0b77de05a87edaab854329fa75df7bf68b29a3def71e985a06e5b5"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "0929bcedde0561d818478961", "attempt_id": "8762ada70218", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:16:36.294743Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 1124, "message_tokens": 1036, "message_utf8_bytes": 2522, "cache_tensor_bytes": null, "serialized_payload_bytes": 2522, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14056.77837494295, "ttft_ms": 14.06566589139402, "decode_ms": 135.07808302529156, "receiver_prefill_ms": 120.89204101357609}, "total_wall_ms": 14328.6693750415, "ttft_ms": 14.06566589139402, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "ZW-73S-2197", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "91aab8fec01151a17508e3b9fa4bc9164d05f5ba8bf8bfc8c47f78922bba670a", "output_hash": "5ab36a19a647e8707dc0e01150dbfc0b49554c16496650f6b32f847c5e5bdf04"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "4ad6cf21b3b39fe9b1096f0a", "attempt_id": "be92ee271183", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:16:36.580300Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 1130, "message_tokens": 1036, "message_utf8_bytes": 2522, "cache_tensor_bytes": null, "serialized_payload_bytes": 2522, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14056.77837494295, "ttft_ms": 13.289707945659757, "decode_ms": 95.14458302874118, "receiver_prefill_ms": 121.742874966003}, "total_wall_ms": 14288.765666889958, "ttft_ms": 13.289707945659757, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "3199125", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "91aab8fec01151a17508e3b9fa4bc9164d05f5ba8bf8bfc8c47f78922bba670a", "output_hash": "cf9a38c1e0ee9fd29d21948a185e19195595b5a7459cd14622cda0879ee474c8"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "786c8567cbaccf73d266918d", "attempt_id": "1d2b21b54178", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:16:36.826125Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 1124, "message_tokens": 1036, "message_utf8_bytes": 2522, "cache_tensor_bytes": null, "serialized_payload_bytes": 2522, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14056.77837494295, "ttft_ms": 13.389083091169596, "decode_ms": 94.58775003440678, "receiver_prefill_ms": 121.54170905705541}, "total_wall_ms": 14288.23750000447, "ttft_ms": 13.389083091169596, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7028925", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "91aab8fec01151a17508e3b9fa4bc9164d05f5ba8bf8bfc8c47f78922bba670a", "output_hash": "e966a398855c104a86bce07742c3ce8108a432926bc704940ecfd07e44596774"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "98008617b90ecd9769390e2c", "attempt_id": "3a5cd2c8de9f", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:16:37.074551Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 1133, "message_tokens": 1036, "message_utf8_bytes": 2522, "cache_tensor_bytes": null, "serialized_payload_bytes": 2522, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14056.77837494295, "ttft_ms": 13.913333998061717, "decode_ms": 80.47712489496917, "receiver_prefill_ms": 120.84470898844302}, "total_wall_ms": 14276.629124884494, "ttft_ms": 13.913333998061717, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Melnak-19", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "91aab8fec01151a17508e3b9fa4bc9164d05f5ba8bf8bfc8c47f78922bba670a", "output_hash": "6553dd0d56d5670d59c0e67ac476769978a10a15b6c08e74fb8332f860a2793c"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "13ceeb781be6bafb8ec2eb2a", "attempt_id": "4e2594c51878", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:16:37.308521Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 1126, "message_tokens": 1036, "message_utf8_bytes": 2522, "cache_tensor_bytes": null, "serialized_payload_bytes": 2522, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14056.77837494295, "ttft_ms": 13.543416978791356, "decode_ms": 93.9698750153184, "receiver_prefill_ms": 120.74858404230326}, "total_wall_ms": 14286.900707986206, "ttft_ms": 13.543416978791356, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1082980", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "91aab8fec01151a17508e3b9fa4bc9164d05f5ba8bf8bfc8c47f78922bba670a", "output_hash": "80f21e206422d6d48e49620c75276d0cc06a056554036f12fbf5f790c2d97006"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a8a060df7b240f49c3a593e5", "attempt_id": "cd2ae036d160", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:16:37.553159Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 1127, "message_tokens": 1036, "message_utf8_bytes": 2522, "cache_tensor_bytes": null, "serialized_payload_bytes": 2522, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14056.77837494295, "ttft_ms": 13.698667054995894, "decode_ms": 13.224458089098334, "receiver_prefill_ms": 121.38183391653001}, "total_wall_ms": 14206.881416030228, "ttft_ms": 13.698667054995894, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "91aab8fec01151a17508e3b9fa4bc9164d05f5ba8bf8bfc8c47f78922bba670a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "9d463e1660b55e8bbb3e5250", "attempt_id": "3cf6551b617f", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:16:37.717235Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 1128, "message_tokens": 1036, "message_utf8_bytes": 2522, "cache_tensor_bytes": null, "serialized_payload_bytes": 2522, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14056.77837494295, "ttft_ms": 13.375082984566689, "decode_ms": 13.229417032562196, "receiver_prefill_ms": 121.42658303491771}, "total_wall_ms": 14206.61079196725, "ttft_ms": 13.375082984566689, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "91aab8fec01151a17508e3b9fa4bc9164d05f5ba8bf8bfc8c47f78922bba670a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "5920484deffb7e0e73b235b3", "attempt_id": "25e02a4a0f55", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:16:51.666319Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 12}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 593, "message_tokens": 501, "message_utf8_bytes": 1307, "cache_tensor_bytes": null, "serialized_payload_bytes": 1307, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13785.146125010215, "ttft_ms": 13.938750023953617, "decode_ms": 97.26329206023365, "receiver_prefill_ms": 70.19079104065895}, "total_wall_ms": 13967.90291601792, "ttft_ms": 13.938750023953617, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8421725", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a3711ad8b78f9d16e212f22b127f57b3007d10ce0e5074fa452fa52879eda42a", "output_hash": "e4949aba3b0b77de05a87edaab854329fa75df7bf68b29a3def71e985a06e5b5"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "52cce99e1448f3db94577be7", "attempt_id": "22dbae1bd91d", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:16:51.864061Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 12}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 593, "message_tokens": 501, "message_utf8_bytes": 1307, "cache_tensor_bytes": null, "serialized_payload_bytes": 1307, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13785.146125010215, "ttft_ms": 14.054834027774632, "decode_ms": 13.974042027257383, "receiver_prefill_ms": 69.5376250660047}, "total_wall_ms": 13884.069166961126, "ttft_ms": 14.054834027774632, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a3711ad8b78f9d16e212f22b127f57b3007d10ce0e5074fa452fa52879eda42a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "3e3c6aef8ef482748409641c", "attempt_id": "1c01ddd2342b", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:16:51.978220Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 12}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 599, "message_tokens": 501, "message_utf8_bytes": 1307, "cache_tensor_bytes": null, "serialized_payload_bytes": 1307, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13785.146125010215, "ttft_ms": 13.51633295416832, "decode_ms": 125.76324993278831, "receiver_prefill_ms": 70.10574999731034}, "total_wall_ms": 13996.554165962152, "ttft_ms": 13.51633295416832, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "3,199,125", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a3711ad8b78f9d16e212f22b127f57b3007d10ce0e5074fa452fa52879eda42a", "output_hash": "f5f7ba5ec2f9b2364ec9e09349c810ade23d11404d04c3b17cea227797de4e79"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "27d829853e809e2ae0d365f2", "attempt_id": "b588bae1afc8", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:16:52.204056Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 12}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 593, "message_tokens": 501, "message_utf8_bytes": 1307, "cache_tensor_bytes": null, "serialized_payload_bytes": 1307, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13785.146125010215, "ttft_ms": 13.65149999037385, "decode_ms": 125.64574996940792, "receiver_prefill_ms": 69.83741698786616}, "total_wall_ms": 13995.610374957323, "ttft_ms": 13.65149999037385, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7,028,925", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a3711ad8b78f9d16e212f22b127f57b3007d10ce0e5074fa452fa52879eda42a", "output_hash": "e83edb3f38a69e604be343e63e2e84442156fcb2bb30f36ba9716575d5ffd57a"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "3b7ab49366c48f5d23513031", "attempt_id": "c11f3d1c82f5", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:16:52.429477Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 12}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 602, "message_tokens": 501, "message_utf8_bytes": 1307, "cache_tensor_bytes": null, "serialized_payload_bytes": 1307, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13785.146125010215, "ttft_ms": 13.62141699064523, "decode_ms": 81.10874996054918, "receiver_prefill_ms": 70.56133402511477}, "total_wall_ms": 13951.72158407513, "ttft_ms": 13.62141699064523, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Melnak-19", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a3711ad8b78f9d16e212f22b127f57b3007d10ce0e5074fa452fa52879eda42a", "output_hash": "6553dd0d56d5670d59c0e67ac476769978a10a15b6c08e74fb8332f860a2793c"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "2c1a993bb76b9a031bff1652", "attempt_id": "6ba88edcb299", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:16:52.610328Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 12}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 595, "message_tokens": 501, "message_utf8_bytes": 1307, "cache_tensor_bytes": null, "serialized_payload_bytes": 1307, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13785.146125010215, "ttft_ms": 13.384416932240129, "decode_ms": 95.17054201569408, "receiver_prefill_ms": 70.22108405362815}, "total_wall_ms": 13965.25412495248, "ttft_ms": 13.384416932240129, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1040475", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a3711ad8b78f9d16e212f22b127f57b3007d10ce0e5074fa452fa52879eda42a", "output_hash": "3c4a6624f259ea52e5db1344f56ad4612327ab9d1256d60fc7008e0790f2c54c"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "65a374a0d7849a4548a40d6c", "attempt_id": "37d7de238c5a", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:16:52.806568Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 12}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 596, "message_tokens": 501, "message_utf8_bytes": 1307, "cache_tensor_bytes": null, "serialized_payload_bytes": 1307, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13785.146125010215, "ttft_ms": 13.668749947100878, "decode_ms": 13.808666029945016, "receiver_prefill_ms": 69.94037493132055}, "total_wall_ms": 13883.858332992531, "ttft_ms": 13.668749947100878, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a3711ad8b78f9d16e212f22b127f57b3007d10ce0e5074fa452fa52879eda42a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "08684b7f81b56a251be498c0", "attempt_id": "8d04544fdbed", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:16:52.919515Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 12}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 597, "message_tokens": 501, "message_utf8_bytes": 1307, "cache_tensor_bytes": null, "serialized_payload_bytes": 1307, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13785.146125010215, "ttft_ms": 13.868374982848763, "decode_ms": 13.78741697408259, "receiver_prefill_ms": 69.94025001768023}, "total_wall_ms": 13884.145584073849, "ttft_ms": 13.868374982848763, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a3711ad8b78f9d16e212f22b127f57b3007d10ce0e5074fa452fa52879eda42a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "ca4843f1a67c22e925a538d5", "attempt_id": "7e5042edaddb", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:17:06.780801Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 1381, "message_tokens": 1290, "message_utf8_bytes": 3375, "cache_tensor_bytes": null, "serialized_payload_bytes": 3375, "side_information_bytes": 0, "fetched_bytes": 1880, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13746.308917063288, "ttft_ms": 13.635165989398956, "decode_ms": 93.98216602858156, "receiver_prefill_ms": 148.7286250339821}, "total_wall_ms": 14004.844584036618, "ttft_ms": 13.635165989398956, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8421725", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "31b9530b3c6f2e8b1619623074fbc99202ee966df3ec864eb0fa204f0846439e", "output_hash": "e4949aba3b0b77de05a87edaab854329fa75df7bf68b29a3def71e985a06e5b5"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "5ee2d067f2ddd055abba890b", "attempt_id": "50da83fef450", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:17:07.054706Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 1333, "message_tokens": 1242, "message_utf8_bytes": 3307, "cache_tensor_bytes": null, "serialized_payload_bytes": 3307, "side_information_bytes": 0, "fetched_bytes": 1811, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13746.308917063288, "ttft_ms": 13.816207996569574, "decode_ms": 13.174165971577168, "receiver_prefill_ms": 143.31066701561213}, "total_wall_ms": 13918.762958957814, "ttft_ms": 13.816207996569574, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "c441f666b0e99a283ed40bfe7fc4a56513412e7c9077c66a6869e4a040106440", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a25c1dc0414e19515a10c51d", "attempt_id": "72cad98305eb", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:17:07.241544Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 1362, "message_tokens": 1265, "message_utf8_bytes": 3284, "cache_tensor_bytes": null, "serialized_payload_bytes": 3284, "side_information_bytes": 0, "fetched_bytes": 1794, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13746.308917063288, "ttft_ms": 13.516457984223962, "decode_ms": 121.0181670030579, "receiver_prefill_ms": 144.48629191610962}, "total_wall_ms": 14027.728583081625, "ttft_ms": 13.516457984223962, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "3,199,125", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "adfccf153cfe68701224e35b19127603b16b3cf28903cbb7cb2bc89fe76262ab", "output_hash": "f5f7ba5ec2f9b2364ec9e09349c810ade23d11404d04c3b17cea227797de4e79"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "9abaecb93a61e460d68b6c95", "attempt_id": "26130e10d4d1", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:17:07.537358Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 1331, "message_tokens": 1240, "message_utf8_bytes": 3318, "cache_tensor_bytes": null, "serialized_payload_bytes": 3318, "side_information_bytes": 0, "fetched_bytes": 1820, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13746.308917063288, "ttft_ms": 13.442500028759241, "decode_ms": 94.21741706319153, "receiver_prefill_ms": 142.3417499754578}, "total_wall_ms": 13998.481750139035, "ttft_ms": 13.442500028759241, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7028925", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ca108979f6d1c716eb394895704f24dbcee9cf0d7a8a35478acedf044fd24bb9", "output_hash": "e966a398855c104a86bce07742c3ce8108a432926bc704940ecfd07e44596774"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "d21781ad6f1c81fe68ff3426", "attempt_id": "f496c7f6ee14", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:17:07.803682Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 1358, "message_tokens": 1258, "message_utf8_bytes": 3288, "cache_tensor_bytes": null, "serialized_payload_bytes": 3288, "side_information_bytes": 0, "fetched_bytes": 1798, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13746.308917063288, "ttft_ms": 13.452792074531317, "decode_ms": 80.40695800445974, "receiver_prefill_ms": 143.8921249937266}, "total_wall_ms": 13986.220375052653, "ttft_ms": 13.452792074531317, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Pragamfel-2", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "7fa4b8ef2a44745ab3fd73539c0a1b74d12fd640bb2c44fbe435556a29a4bd51", "output_hash": "f0df32e3152d54160c1f3bdeb20ad87cbc7e1efcd3ea28c84ac1d2be0ed97c7a"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "5a766071b417309ea9cecb92", "attempt_id": "7d0021827521", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:17:08.060688Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 1383, "message_tokens": 1290, "message_utf8_bytes": 3375, "cache_tensor_bytes": null, "serialized_payload_bytes": 3375, "side_information_bytes": 0, "fetched_bytes": 1880, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13746.308917063288, "ttft_ms": 13.370375032536685, "decode_ms": 94.08504201564938, "receiver_prefill_ms": 148.23129191063344}, "total_wall_ms": 14004.162917030044, "ttft_ms": 13.370375032536685, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5877577", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "31b9530b3c6f2e8b1619623074fbc99202ee966df3ec864eb0fa204f0846439e", "output_hash": "74089ccbf202faac2fd195d97f9fbbe2c5c231a3cb9544a51af9e80e4667f859"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "02c5afc262f86d4d863ac4f0", "attempt_id": "5a374979af24", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:17:08.332721Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 1336, "message_tokens": 1242, "message_utf8_bytes": 3307, "cache_tensor_bytes": null, "serialized_payload_bytes": 3307, "side_information_bytes": 0, "fetched_bytes": 1811, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13746.308917063288, "ttft_ms": 13.357916031964123, "decode_ms": 13.45525006763637, "receiver_prefill_ms": 143.25491699855775}, "total_wall_ms": 13918.514667078853, "ttft_ms": 13.357916031964123, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "c441f666b0e99a283ed40bfe7fc4a56513412e7c9077c66a6869e4a040106440", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "1ca122aa615ae59feacd2c9b", "attempt_id": "3e0bad8148fd", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:17:08.519120Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 1337, "message_tokens": 1242, "message_utf8_bytes": 3307, "cache_tensor_bytes": null, "serialized_payload_bytes": 3307, "side_information_bytes": 0, "fetched_bytes": 1811, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13746.308917063288, "ttft_ms": 13.911540969274938, "decode_ms": 13.38800007943064, "receiver_prefill_ms": 147.26708398666233}, "total_wall_ms": 13923.029542085715, "ttft_ms": 13.911540969274938, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "c441f666b0e99a283ed40bfe7fc4a56513412e7c9077c66a6869e4a040106440", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "84aa47fca54d101f1a0f48c1", "attempt_id": "eae2c611aa63", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:17:08.994103Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 84676608}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 45, "message_tokens": null, "message_utf8_bytes": 8078, "cache_tensor_bytes": 84676608, "serialized_payload_bytes": 84676608, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 284.18025001883507, "ttft_ms": 33.907667035236955, "decode_ms": 98.38133305311203, "receiver_prefill_ms": 0.0}, "total_wall_ms": 420.60754203703254, "ttft_ms": 33.907667035236955, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8421725", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "e4949aba3b0b77de05a87edaab854329fa75df7bf68b29a3def71e985a06e5b5"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "dfe8d67b3153dd999a1c14d2", "attempt_id": "e028b1183dc3", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:17:09.145119Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 84676608}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 45, "message_tokens": null, "message_utf8_bytes": 8078, "cache_tensor_bytes": 84676608, "serialized_payload_bytes": 84676608, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 284.18025001883507, "ttft_ms": 28.543333057314157, "decode_ms": 138.59070895705372, "receiver_prefill_ms": 0.0}, "total_wall_ms": 455.1207080949098, "ttft_ms": 28.543333057314157, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "ZW-73S-2197", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "5ab36a19a647e8707dc0e01150dbfc0b49554c16496650f6b32f847c5e5bdf04"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "6100b52e04e687094daf9da9", "attempt_id": "f964e8568c7a", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:17:09.330330Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 84676608}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 51, "message_tokens": null, "message_utf8_bytes": 8078, "cache_tensor_bytes": 84676608, "serialized_payload_bytes": 84676608, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 284.18025001883507, "ttft_ms": 27.388166985474527, "decode_ms": 97.1458749845624, "receiver_prefill_ms": 0.0}, "total_wall_ms": 412.55391703452915, "ttft_ms": 27.388166985474527, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "3199125", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "cf9a38c1e0ee9fd29d21948a185e19195595b5a7459cd14622cda0879ee474c8"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "f872ba420801d12d30fb9823", "attempt_id": "fab8f59fe55f", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:17:09.473098Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 84676608}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 45, "message_tokens": null, "message_utf8_bytes": 8078, "cache_tensor_bytes": 84676608, "serialized_payload_bytes": 84676608, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 284.18025001883507, "ttft_ms": 28.237750055268407, "decode_ms": 98.0200000340119, "receiver_prefill_ms": 0.0}, "total_wall_ms": 414.2843750305474, "ttft_ms": 28.237750055268407, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8377480", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "eedcb3ec2bdef623b3fb1b0d8a045a8680babcb6cf46ed7f48d4d0c430347e5c"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "dd6bd09f8898662208eb19a1", "attempt_id": "7b303596aca8", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:17:09.617515Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 84676608}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 54, "message_tokens": null, "message_utf8_bytes": 8078, "cache_tensor_bytes": 84676608, "serialized_payload_bytes": 84676608, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 284.18025001883507, "ttft_ms": 28.532749973237514, "decode_ms": 83.0171249108389, "receiver_prefill_ms": 0.0}, "total_wall_ms": 399.41604202613235, "ttft_ms": 28.532749973237514, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Pragamfel-2", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "f0df32e3152d54160c1f3bdeb20ad87cbc7e1efcd3ea28c84ac1d2be0ed97c7a"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "6708ec1034daf8ad0a3d9baf", "attempt_id": "2b6688a6c48a", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:17:09.746966Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 84676608}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 8078, "cache_tensor_bytes": 84676608, "serialized_payload_bytes": 84676608, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 284.18025001883507, "ttft_ms": 27.450083987787366, "decode_ms": 98.85958302766085, "receiver_prefill_ms": 0.0}, "total_wall_ms": 414.22162507660687, "ttft_ms": 27.450083987787366, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4959747", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "db4d8d3b3740ba5c0eb24ddd8ecd4e41b0fafe597f33028e14a89f85f4857a80"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "f792797e5bd30946e9b7f213", "attempt_id": "6c3e5cfbf085", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:17:09.891582Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 84676608}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 48, "message_tokens": null, "message_utf8_bytes": 8078, "cache_tensor_bytes": 84676608, "serialized_payload_bytes": 84676608, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 284.18025001883507, "ttft_ms": 28.400583076290786, "decode_ms": 14.186916989274323, "receiver_prefill_ms": 0.0}, "total_wall_ms": 330.6900840252638, "ttft_ms": 28.400583076290786, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "4d82ac59e085e3c31e6da1ae", "attempt_id": "0f327bcc616c", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:17:09.952364Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 84676608}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 49, "message_tokens": null, "message_utf8_bytes": 8078, "cache_tensor_bytes": 84676608, "serialized_payload_bytes": 84676608, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 284.18025001883507, "ttft_ms": 28.007791028358042, "decode_ms": 13.797541032545269, "receiver_prefill_ms": 0.0}, "total_wall_ms": 329.9325000261888, "ttft_ms": 28.007791028358042, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "281b40dbd29f5c91a1cd1990", "attempt_id": "4cf51bbf0055", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:17:10.012527Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 2420, "message_tokens": 2333, "message_utf8_bytes": 8608, "cache_tensor_bytes": null, "serialized_payload_bytes": 8608, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.025499961338937283, "ttft_ms": 13.546000001952052, "decode_ms": 94.90791603457183, "receiver_prefill_ms": 265.00058302190155}, "total_wall_ms": 377.16045789420605, "ttft_ms": 13.546000001952052, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8421725", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "0e8fbd589c17bb7645075de4f62e683affc19600b6fa6404da97d926e59d5ca5", "output_hash": "e4949aba3b0b77de05a87edaab854329fa75df7bf68b29a3def71e985a06e5b5"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "9465e66836cedddbd1bc2e60", "attempt_id": "127321081397", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:17:10.403318Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 2420, "message_tokens": 2333, "message_utf8_bytes": 8608, "cache_tensor_bytes": null, "serialized_payload_bytes": 8608, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.025499961338937283, "ttft_ms": 13.616833020932972, "decode_ms": 136.20125001762062, "receiver_prefill_ms": 266.89954206813127}, "total_wall_ms": 420.49412487540394, "ttft_ms": 13.616833020932972, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "ZW-73S-2197", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "0e8fbd589c17bb7645075de4f62e683affc19600b6fa6404da97d926e59d5ca5", "output_hash": "5ab36a19a647e8707dc0e01150dbfc0b49554c16496650f6b32f847c5e5bdf04"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "8f07f173a9404aed04c95352", "attempt_id": "617c409345b3", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:17:10.838045Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 2426, "message_tokens": 2333, "message_utf8_bytes": 8608, "cache_tensor_bytes": null, "serialized_payload_bytes": 8608, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.025499961338937283, "ttft_ms": 13.589999987743795, "decode_ms": 95.14200000558048, "receiver_prefill_ms": 265.1636250084266}, "total_wall_ms": 377.7361249085516, "ttft_ms": 13.589999987743795, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "3199125", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "0e8fbd589c17bb7645075de4f62e683affc19600b6fa6404da97d926e59d5ca5", "output_hash": "cf9a38c1e0ee9fd29d21948a185e19195595b5a7459cd14622cda0879ee474c8"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "7c8985dc025d65ad661eecb5", "attempt_id": "a43f58656a2b", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:17:11.230129Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 2420, "message_tokens": 2333, "message_utf8_bytes": 8608, "cache_tensor_bytes": null, "serialized_payload_bytes": 8608, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.025499961338937283, "ttft_ms": 13.828792027197778, "decode_ms": 95.26374994311482, "receiver_prefill_ms": 277.66129199881107}, "total_wall_ms": 390.4120830120519, "ttft_ms": 13.828792027197778, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7791055", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "0e8fbd589c17bb7645075de4f62e683affc19600b6fa6404da97d926e59d5ca5", "output_hash": "b226e01df2578a21c258f145c74ce4aae110dcd20ba1e5c4b585535f188744a7"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "29f15fdba9520f470625975c", "attempt_id": "265f63930b0a", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:17:11.635111Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 2429, "message_tokens": 2333, "message_utf8_bytes": 8608, "cache_tensor_bytes": null, "serialized_payload_bytes": 8608, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.025499961338937283, "ttft_ms": 14.758917037397623, "decode_ms": 81.99666708242148, "receiver_prefill_ms": 266.9758329866454}, "total_wall_ms": 367.53208306618035, "ttft_ms": 14.758917037397623, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Pragamfel-2", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "0e8fbd589c17bb7645075de4f62e683affc19600b6fa6404da97d926e59d5ca5", "output_hash": "f0df32e3152d54160c1f3bdeb20ad87cbc7e1efcd3ea28c84ac1d2be0ed97c7a"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b312bb58dccaf208c1756aed", "attempt_id": "a799f56d9a3e", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:17:12.017337Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 2422, "message_tokens": 2333, "message_utf8_bytes": 8608, "cache_tensor_bytes": null, "serialized_payload_bytes": 8608, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.025499961338937283, "ttft_ms": 13.921083067543805, "decode_ms": 94.8648750782013, "receiver_prefill_ms": 269.2109999479726}, "total_wall_ms": 381.88608293421566, "ttft_ms": 13.921083067543805, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5204729", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "0e8fbd589c17bb7645075de4f62e683affc19600b6fa6404da97d926e59d5ca5", "output_hash": "3ca96c58a9baaf41ff69614fd2be180a2bde4b00e634ab230489c3c26d4b2308"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "81f08056d29863715fbe9c2a", "attempt_id": "b93ab10a32f7", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:17:12.413702Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 2423, "message_tokens": 2333, "message_utf8_bytes": 8608, "cache_tensor_bytes": null, "serialized_payload_bytes": 8608, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.025499961338937283, "ttft_ms": 13.778832973912358, "decode_ms": 13.75499996356666, "receiver_prefill_ms": 270.1174169778824}, "total_wall_ms": 301.2994579039514, "ttft_ms": 13.778832973912358, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "0e8fbd589c17bb7645075de4f62e683affc19600b6fa6404da97d926e59d5ca5", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a9c2fabe7c5cf4fb5536cdbd", "attempt_id": "a442d8e757a8", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:17:12.729339Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 2424, "message_tokens": 2333, "message_utf8_bytes": 8608, "cache_tensor_bytes": null, "serialized_payload_bytes": 8608, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.025499961338937283, "ttft_ms": 13.794124941341579, "decode_ms": 13.820749940350652, "receiver_prefill_ms": 286.43704194109887}, "total_wall_ms": 317.86133302375674, "ttft_ms": 13.794124941341579, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "0e8fbd589c17bb7645075de4f62e683affc19600b6fa6404da97d926e59d5ca5", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "3afa836b0994967a7ef28992", "attempt_id": "5cf971508be5", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:17:34.970962Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1609}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 1722, "message_tokens": 1634, "message_utf8_bytes": 4861, "cache_tensor_bytes": null, "serialized_payload_bytes": 4861, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 21909.953208989464, "ttft_ms": 13.660582946613431, "decode_ms": 95.40779201779515, "receiver_prefill_ms": 180.86783296894282}, "total_wall_ms": 22207.72633398883, "ttft_ms": 13.660582946613431, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8421725", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a69d5142a71207d880df131357924c3d0448a9fbcafb1ba43a3d76c4009e794c", "output_hash": "e4949aba3b0b77de05a87edaab854329fa75df7bf68b29a3def71e985a06e5b5"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "7b2461f317a1bb601870dfc8", "attempt_id": "62a3d6ddfa84", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:17:35.283837Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1609}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 1722, "message_tokens": 1634, "message_utf8_bytes": 4861, "cache_tensor_bytes": null, "serialized_payload_bytes": 4861, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 21909.953208989464, "ttft_ms": 13.546250062063336, "decode_ms": 134.4941669376567, "receiver_prefill_ms": 179.22641697805375}, "total_wall_ms": 22239.824416930787, "ttft_ms": 13.546250062063336, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "ZW-73S-2197", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a69d5142a71207d880df131357924c3d0448a9fbcafb1ba43a3d76c4009e794c", "output_hash": "5ab36a19a647e8707dc0e01150dbfc0b49554c16496650f6b32f847c5e5bdf04"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "7a8f4b4b312971cbfecfa158", "attempt_id": "e99293d11a3b", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:17:35.628396Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1609}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 1728, "message_tokens": 1634, "message_utf8_bytes": 4861, "cache_tensor_bytes": null, "serialized_payload_bytes": 4861, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 21909.953208989464, "ttft_ms": 13.34983401466161, "decode_ms": 93.78204110544175, "receiver_prefill_ms": 181.09191698022187}, "total_wall_ms": 22200.786001048982, "ttft_ms": 13.34983401466161, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "3199125", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a69d5142a71207d880df131357924c3d0448a9fbcafb1ba43a3d76c4009e794c", "output_hash": "cf9a38c1e0ee9fd29d21948a185e19195595b5a7459cd14622cda0879ee474c8"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "22e162b8e6f10d6a657416c6", "attempt_id": "f50284a8ae5e", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:17:35.933565Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1609}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 1722, "message_tokens": 1634, "message_utf8_bytes": 4861, "cache_tensor_bytes": null, "serialized_payload_bytes": 4861, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 21909.953208989464, "ttft_ms": 13.489500037394464, "decode_ms": 120.45166606549174, "receiver_prefill_ms": 179.3373340042308}, "total_wall_ms": 22225.852124975063, "ttft_ms": 13.489500037394464, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7,791,055", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a69d5142a71207d880df131357924c3d0448a9fbcafb1ba43a3d76c4009e794c", "output_hash": "a7a9e7d4ae8669152180c3b8a16087d50f29ecb3daa859b35b9aef6996508884"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "5788718a79dc9fc4cbbc7015", "attempt_id": "a3f6b19d79a2", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:17:36.263589Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1609}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 1731, "message_tokens": 1634, "message_utf8_bytes": 4861, "cache_tensor_bytes": null, "serialized_payload_bytes": 4861, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 21909.953208989464, "ttft_ms": 13.3819580078125, "decode_ms": 94.92125001270324, "receiver_prefill_ms": 182.75008304044604}, "total_wall_ms": 22203.642750973813, "ttft_ms": 13.3819580078125, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Lithzentek-13", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a69d5142a71207d880df131357924c3d0448a9fbcafb1ba43a3d76c4009e794c", "output_hash": "155ddc5175ca295a0cb2ea7f3b9ea2adeffb2d4da429843370cd08b36d787ca0"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "9a33ea12f3b1b3eeb7bb089b", "attempt_id": "e78a446efecc", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:17:36.571493Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1609}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 1724, "message_tokens": 1634, "message_utf8_bytes": 4861, "cache_tensor_bytes": null, "serialized_payload_bytes": 4861, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 21909.953208989464, "ttft_ms": 13.849499984644353, "decode_ms": 95.27249995153397, "receiver_prefill_ms": 194.93816699832678}, "total_wall_ms": 22216.63683396764, "ttft_ms": 13.849499984644353, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5874507", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a69d5142a71207d880df131357924c3d0448a9fbcafb1ba43a3d76c4009e794c", "output_hash": "519d339a5ce07dece5324ee1a84ccbf0a076058cb314823a49931a4460c99261"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "234598f59ddeb5108b69c2e8", "attempt_id": "0bfd721ed272", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:17:36.892749Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1609}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 1725, "message_tokens": 1634, "message_utf8_bytes": 4861, "cache_tensor_bytes": null, "serialized_payload_bytes": 4861, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 21909.953208989464, "ttft_ms": 13.6943330289796, "decode_ms": 13.596583041362464, "receiver_prefill_ms": 191.02037500124425}, "total_wall_ms": 22130.943708936684, "ttft_ms": 13.6943330289796, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a69d5142a71207d880df131357924c3d0448a9fbcafb1ba43a3d76c4009e794c", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "17731610c0f3fbe07c37d384", "attempt_id": "508cf0d750af", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:17:37.127978Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1609}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 1726, "message_tokens": 1634, "message_utf8_bytes": 4861, "cache_tensor_bytes": null, "serialized_payload_bytes": 4861, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 21909.953208989464, "ttft_ms": 13.669083011336625, "decode_ms": 13.552832999266684, "receiver_prefill_ms": 200.66079206299037}, "total_wall_ms": 22140.460584079847, "ttft_ms": 13.669083011336625, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a69d5142a71207d880df131357924c3d0448a9fbcafb1ba43a3d76c4009e794c", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "2f32df36a3813826b468dd49", "attempt_id": "efe8ac55da23", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:18:00.736274Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 25}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 1056, "message_tokens": 964, "message_utf8_bytes": 2594, "cache_tensor_bytes": null, "serialized_payload_bytes": 2594, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 23363.320708973333, "ttft_ms": 14.230792061425745, "decode_ms": 13.7020000256598, "receiver_prefill_ms": 483.54387492872775}, "total_wall_ms": 23876.79362599738, "ttft_ms": 14.230792061425745, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ecc73f87bbb6ce54e39c88e19356ced400a8ccba547bcd4993b3fbcb6e41f728", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "7d7a8233f01bbe25777cb08d", "attempt_id": "21e098d7e6b1", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:18:01.270626Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 25}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 1056, "message_tokens": 964, "message_utf8_bytes": 2594, "cache_tensor_bytes": null, "serialized_payload_bytes": 2594, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 23363.320708973333, "ttft_ms": 13.630958972498775, "decode_ms": 13.655124930664897, "receiver_prefill_ms": 113.1226250436157}, "total_wall_ms": 23505.532208946534, "ttft_ms": 13.630958972498775, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ecc73f87bbb6ce54e39c88e19356ced400a8ccba547bcd4993b3fbcb6e41f728", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "4ebb59b94b9ba9daf1247c1b", "attempt_id": "5e92ef9b29f7", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:18:01.428928Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 25}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 1062, "message_tokens": 964, "message_utf8_bytes": 2594, "cache_tensor_bytes": null, "serialized_payload_bytes": 2594, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 23363.320708973333, "ttft_ms": 61.36379099916667, "decode_ms": 98.11770799569786, "receiver_prefill_ms": 1740.5529170064256}, "total_wall_ms": 25265.115125919692, "ttft_ms": 61.36379099916667, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "3199125", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ecc73f87bbb6ce54e39c88e19356ced400a8ccba547bcd4993b3fbcb6e41f728", "output_hash": "cf9a38c1e0ee9fd29d21948a185e19195595b5a7459cd14622cda0879ee474c8"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "784edd03af11e5de69e6ee73", "attempt_id": "1d1690232adc", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:18:03.345334Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 25}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 1056, "message_tokens": 964, "message_utf8_bytes": 2594, "cache_tensor_bytes": null, "serialized_payload_bytes": 2594, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 23363.320708973333, "ttft_ms": 14.025290962308645, "decode_ms": 97.19854197464883, "receiver_prefill_ms": 157.10595797281712}, "total_wall_ms": 23633.60974995885, "ttft_ms": 14.025290962308645, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7028925", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ecc73f87bbb6ce54e39c88e19356ced400a8ccba547bcd4993b3fbcb6e41f728", "output_hash": "e966a398855c104a86bce07742c3ce8108a432926bc704940ecfd07e44596774"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "1066a4b8b202261cea1aea75", "attempt_id": "625cb60060e6", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:18:03.637594Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 25}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 1065, "message_tokens": 964, "message_utf8_bytes": 2594, "cache_tensor_bytes": null, "serialized_payload_bytes": 2594, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 23363.320708973333, "ttft_ms": 13.783250004053116, "decode_ms": 83.62762501928955, "receiver_prefill_ms": 115.1902920100838}, "total_wall_ms": 23577.777251019143, "ttft_ms": 13.783250004053116, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Melnak-19", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ecc73f87bbb6ce54e39c88e19356ced400a8ccba547bcd4993b3fbcb6e41f728", "output_hash": "6553dd0d56d5670d59c0e67ac476769978a10a15b6c08e74fb8332f860a2793c"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "7a98d75e7338808cce07a872", "attempt_id": "1bddec1eeac0", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:18:03.866973Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 25}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 1058, "message_tokens": 964, "message_utf8_bytes": 2594, "cache_tensor_bytes": null, "serialized_payload_bytes": 2594, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 23363.320708973333, "ttft_ms": 13.427959056571126, "decode_ms": 93.98462506942451, "receiver_prefill_ms": 115.31820800155401}, "total_wall_ms": 23587.846667971462, "ttft_ms": 13.427959056571126, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1048737", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ecc73f87bbb6ce54e39c88e19356ced400a8ccba547bcd4993b3fbcb6e41f728", "output_hash": "c55fee283c3bbce60d07b1a4cb4e067950de31ac134360ae722241154bfc878e"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "2a8247d45c3ce031b8f26dd0", "attempt_id": "9e87df63efe4", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:18:04.106796Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 25}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 1059, "message_tokens": 964, "message_utf8_bytes": 2594, "cache_tensor_bytes": null, "serialized_payload_bytes": 2594, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 23363.320708973333, "ttft_ms": 13.539084000512958, "decode_ms": 13.242790941148996, "receiver_prefill_ms": 115.09291699621826}, "total_wall_ms": 23506.868668016978, "ttft_ms": 13.539084000512958, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ecc73f87bbb6ce54e39c88e19356ced400a8ccba547bcd4993b3fbcb6e41f728", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "e516816dee5ceec6b552e24c", "attempt_id": "5c0faa375b18", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:18:04.264792Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 25}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 1060, "message_tokens": 964, "message_utf8_bytes": 2594, "cache_tensor_bytes": null, "serialized_payload_bytes": 2594, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 23363.320708973333, "ttft_ms": 13.839749968610704, "decode_ms": 13.771665981039405, "receiver_prefill_ms": 115.54958298802376}, "total_wall_ms": 23508.234000997618, "ttft_ms": 13.839749968610704, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ecc73f87bbb6ce54e39c88e19356ced400a8ccba547bcd4993b3fbcb6e41f728", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "fca58ba1da3ea61372714bd7", "attempt_id": "ac6b58fa4a69", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:18:27.973087Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 2522, "message_tokens": 2431, "message_utf8_bytes": 6407, "cache_tensor_bytes": null, "serialized_payload_bytes": 6407, "side_information_bytes": 0, "fetched_bytes": 3450, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 23549.433334032074, "ttft_ms": 13.693250017240644, "decode_ms": 94.4837910356, "receiver_prefill_ms": 331.5962089691311}, "total_wall_ms": 23995.44779304415, "ttft_ms": 13.693250017240644, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8421725", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "91d29cbbc5005e742133d1a007a41520607f60e081f6b4c3631792d5fe586afc", "output_hash": "e4949aba3b0b77de05a87edaab854329fa75df7bf68b29a3def71e985a06e5b5"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "3c67d9b5ed0dd3491d69fb92", "attempt_id": "9335003dbbe0", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:18:28.433591Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 2448, "message_tokens": 2357, "message_utf8_bytes": 6205, "cache_tensor_bytes": null, "serialized_payload_bytes": 6205, "side_information_bytes": 0, "fetched_bytes": 3257, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 23549.433334032074, "ttft_ms": 13.528415933251381, "decode_ms": 13.365959050133824, "receiver_prefill_ms": 269.0090839751065}, "total_wall_ms": 23849.694793112576, "ttft_ms": 13.528415933251381, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "df1b21bf4be579607d70771a51d5e859f299634a1b0656881617f6def8d07206", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "802715f64773c152e8d99dc2", "attempt_id": "962fbfab5592", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:18:28.747815Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 2534, "message_tokens": 2437, "message_utf8_bytes": 6386, "cache_tensor_bytes": null, "serialized_payload_bytes": 6386, "side_information_bytes": 0, "fetched_bytes": 3431, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 23549.433334032074, "ttft_ms": 13.701707939617336, "decode_ms": 96.01462504360825, "receiver_prefill_ms": 273.01879203878343}, "total_wall_ms": 23935.54724997375, "ttft_ms": 13.701707939617336, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "3199125", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ed68d933d17f47479d3510045946144c29eee1c063c2a0ba1de297fb0e35204f", "output_hash": "cf9a38c1e0ee9fd29d21948a185e19195595b5a7459cd14622cda0879ee474c8"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c284a6c47ec1eb25bab12e0c", "attempt_id": "e2f97fbb38c4", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:18:29.147682Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 2528, "message_tokens": 2437, "message_utf8_bytes": 6386, "cache_tensor_bytes": null, "serialized_payload_bytes": 6386, "side_information_bytes": 0, "fetched_bytes": 3431, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 23549.433334032074, "ttft_ms": 13.977332971990108, "decode_ms": 95.07495898287743, "receiver_prefill_ms": 523.9327499875799}, "total_wall_ms": 24185.78229297418, "ttft_ms": 13.977332971990108, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7028925", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ed68d933d17f47479d3510045946144c29eee1c063c2a0ba1de297fb0e35204f", "output_hash": "e966a398855c104a86bce07742c3ce8108a432926bc704940ecfd07e44596774"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "52673f48845de80c5b6395d2", "attempt_id": "5368142fc929", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:18:29.800195Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 2421, "message_tokens": 2321, "message_utf8_bytes": 6209, "cache_tensor_bytes": null, "serialized_payload_bytes": 6209, "side_information_bytes": 0, "fetched_bytes": 3283, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 23549.433334032074, "ttft_ms": 13.831999967806041, "decode_ms": 81.9044589297846, "receiver_prefill_ms": 279.01533304248005}, "total_wall_ms": 23928.430041996762, "ttft_ms": 13.831999967806041, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Pragamfel-2", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "12b1ba31e386092ca4f1beb8801b6a5368930bc501fafa6d3d697288dc414555", "output_hash": "f0df32e3152d54160c1f3bdeb20ad87cbc7e1efcd3ea28c84ac1d2be0ed97c7a"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "20621099a1d3a6ccc8c55f52", "attempt_id": "5c6e5ef18d6d", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:18:30.194359Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 2530, "message_tokens": 2437, "message_utf8_bytes": 6386, "cache_tensor_bytes": null, "serialized_payload_bytes": 6386, "side_information_bytes": 0, "fetched_bytes": 3431, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 23549.433334032074, "ttft_ms": 14.120916952379048, "decode_ms": 96.49595804512501, "receiver_prefill_ms": 294.13033300079405}, "total_wall_ms": 23957.548875943758, "ttft_ms": 14.120916952379048, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5427027", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "1f48d4f354426860827ac5a8658524585cf4420df4242fafbf5d9010dfa7a790", "output_hash": "fc9ef1a090f1b0fd0bd40e70074756ae642efaf42e523a3e1265b33978dc807f"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "d5e830d08093c751e0ab6864", "attempt_id": "262cba52fff5", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:18:30.616840Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 2506, "message_tokens": 2412, "message_utf8_bytes": 6403, "cache_tensor_bytes": null, "serialized_payload_bytes": 6403, "side_information_bytes": 0, "fetched_bytes": 3422, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 23549.433334032074, "ttft_ms": 13.84245790541172, "decode_ms": 13.601666898466647, "receiver_prefill_ms": 281.1926659196615}, "total_wall_ms": 23861.385751049966, "ttft_ms": 13.84245790541172, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "dd598be84b9e834e98c251fea70983ee816ae98042739b38af4119b885d60519", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b6e05b5e52b77b6332eb8450", "attempt_id": "ec2ec0bcf3f3", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:18:30.943544Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 2526, "message_tokens": 2431, "message_utf8_bytes": 6385, "cache_tensor_bytes": null, "serialized_payload_bytes": 6385, "side_information_bytes": 0, "fetched_bytes": 3436, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 23549.433334032074, "ttft_ms": 13.98629101458937, "decode_ms": 13.900207937695086, "receiver_prefill_ms": 302.62129206676036}, "total_wall_ms": 23883.30545905046, "ttft_ms": 13.98629101458937, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "820398775e70dc2251056ccda3fa05ff477d1a84abc7162e336b892fc65eacad", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "06963f4e47a906a465b9b59d", "attempt_id": "871b656478fc", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:18:32.631953Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 91754496}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 45, "message_tokens": null, "message_utf8_bytes": 8554, "cache_tensor_bytes": 91754496, "serialized_payload_bytes": 91754496, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 1340.5482079833746, "ttft_ms": 29.50841700658202, "decode_ms": 98.52770809084177, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1472.0369999995455, "ttft_ms": 29.50841700658202, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8421725", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "e4949aba3b0b77de05a87edaab854329fa75df7bf68b29a3def71e985a06e5b5"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "cd937a0cfc13d84aa2e98af0", "attempt_id": "240c49004cb0", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:18:32.777653Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 91754496}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 45, "message_tokens": null, "message_utf8_bytes": 8554, "cache_tensor_bytes": 91754496, "serialized_payload_bytes": 91754496, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 1340.5482079833746, "ttft_ms": 29.342459049075842, "decode_ms": 139.08849994186312, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1512.0296250097454, "ttft_ms": 29.342459049075842, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "ZW-73S-2197", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "5ab36a19a647e8707dc0e01150dbfc0b49554c16496650f6b32f847c5e5bdf04"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "8ff5bbb67b9d4703d84d22ef", "attempt_id": "30228d80d861", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:18:32.963434Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 91754496}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 51, "message_tokens": null, "message_utf8_bytes": 8554, "cache_tensor_bytes": 91754496, "serialized_payload_bytes": 91754496, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 1340.5482079833746, "ttft_ms": 28.737875050865114, "decode_ms": 125.42408297304064, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1497.3298739641905, "ttft_ms": 28.737875050865114, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8,073,060", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "657104c35541b20cc2676fb4e100c4f27aa55e265abfb2fce5c110b9ab6d8f9f"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "8f32057d188248f87e575343", "attempt_id": "26def2baf8d7", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:18:33.134288Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 91754496}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 45, "message_tokens": null, "message_utf8_bytes": 8554, "cache_tensor_bytes": 91754496, "serialized_payload_bytes": 91754496, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 1340.5482079833746, "ttft_ms": 29.116833000443876, "decode_ms": 125.32174994703382, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1497.7907079737633, "ttft_ms": 29.116833000443876, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8,377,480", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "ec4c68b327a2d2427ed21821415a5f9ddf87134bd9f654bdd69c7d1301febcb0"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "8f9a9c68cd3d4cec1eb547d3", "attempt_id": "d0890cd18c78", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:18:33.306088Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 91754496}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 54, "message_tokens": null, "message_utf8_bytes": 8554, "cache_tensor_bytes": 91754496, "serialized_payload_bytes": 91754496, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 1340.5482079833746, "ttft_ms": 30.020499951206148, "decode_ms": 83.58391700312495, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1457.0481668924913, "ttft_ms": 30.020499951206148, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Melnak-19", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "6553dd0d56d5670d59c0e67ac476769978a10a15b6c08e74fb8332f860a2793c"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "94c4fa16877c3d01ed6ec96c", "attempt_id": "0c8379284e42", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:18:33.436618Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 91754496}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 8554, "cache_tensor_bytes": 91754496, "serialized_payload_bytes": 91754496, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 1340.5482079833746, "ttft_ms": 28.012041933834553, "decode_ms": 111.41083296388388, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1483.2182910759002, "ttft_ms": 28.012041933834553, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "10979777", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "9ee6cb0ebc96f4a3e40aed41c2761d6dccdb8d56f175ee9f54474d5f1aee19bf"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "5d5c83a2357cae93a790f4a3", "attempt_id": "eeb5c320950f", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:18:33.593497Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 91754496}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 48, "message_tokens": null, "message_utf8_bytes": 8554, "cache_tensor_bytes": 91754496, "serialized_payload_bytes": 91754496, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 1340.5482079833746, "ttft_ms": 29.514709021896124, "decode_ms": 14.134124969132245, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1386.2489159218967, "ttft_ms": 29.514709021896124, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c0f9756bcfc3be7b87c1bc8b", "attempt_id": "f3a3ef1a0c2c", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:18:33.653693Z", "task_family": "synthetic", "document_id": "syn_03015_2000", "document_hash": "71a7284b9bc3f681d69d91d18b9e8d45a4dda486cd14f4145395786ac6a68a89", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03015_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 91754496}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1922, "query_tokens": 49, "message_tokens": null, "message_utf8_bytes": 8554, "cache_tensor_bytes": 91754496, "serialized_payload_bytes": 91754496, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 1340.5482079833746, "ttft_ms": 28.868083958514035, "decode_ms": 13.798499945551157, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1386.2264170311391, "ttft_ms": 28.868083958514035, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lithzentek-13", "bogus_amount": "7,791,055", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03015_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b6286b4252207b29e9954397", "attempt_id": "09c73a2f158f", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:18:34.659587Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 6184, "message_tokens": 6098, "message_utf8_bytes": 30224, "cache_tensor_bytes": null, "serialized_payload_bytes": 30224, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.29495800845324993, "ttft_ms": 14.38983401749283, "decode_ms": 124.72341593820602, "receiver_prefill_ms": 733.5297499084845}, "total_wall_ms": 883.5837909718975, "ttft_ms": 14.38983401749283, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7,755,450", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S1"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "43187207d8fddf248ff8736f810b18469b7564f5d39160f89201d1929ee25558", "output_hash": "50ce1bf5a6a1535bae98114196498afa876eadf9e5680cdc8087d8bc8f71357e"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "d335b5ecb9ab10168db85653", "attempt_id": "9ead718ff9ee", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:18:35.559500Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 6186, "message_tokens": 6098, "message_utf8_bytes": 30224, "cache_tensor_bytes": null, "serialized_payload_bytes": 30224, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.29495800845324993, "ttft_ms": 13.966333004646003, "decode_ms": 138.93041596747935, "receiver_prefill_ms": 722.5928750121966}, "total_wall_ms": 885.4032920207828, "ttft_ms": 13.966333004646003, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "RY-99D-9177", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S1"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "43187207d8fddf248ff8736f810b18469b7564f5d39160f89201d1929ee25558", "output_hash": "abc5e1adc6533afa7ee1683a2195252eb090b08b52392c8150ccf6a387b14df6"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a4b8ab9371d29b56e187c2aa", "attempt_id": "62b6626a0834", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:18:36.461273Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 6192, "message_tokens": 6098, "message_utf8_bytes": 30224, "cache_tensor_bytes": null, "serialized_payload_bytes": 30224, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.29495800845324993, "ttft_ms": 13.761166017502546, "decode_ms": 95.60975001659244, "receiver_prefill_ms": 721.7479579849169}, "total_wall_ms": 840.8014579908922, "ttft_ms": 13.761166017502546, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7430110", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S1"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "43187207d8fddf248ff8736f810b18469b7564f5d39160f89201d1929ee25558", "output_hash": "3781b0f7175e64bec5fc4e06e8c55b65284738b62891371384a1ee1dc8c3cfe9"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "e366c0283aa94161a91f08b7", "attempt_id": "a93b8497aed1", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:18:37.318386Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 6185, "message_tokens": 6098, "message_utf8_bytes": 30224, "cache_tensor_bytes": null, "serialized_payload_bytes": 30224, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.29495800845324993, "ttft_ms": 13.746500015258789, "decode_ms": 95.41729092597961, "receiver_prefill_ms": 720.7744580227882}, "total_wall_ms": 839.7995420964435, "ttft_ms": 13.746500015258789, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7430110", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S1"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "43187207d8fddf248ff8736f810b18469b7564f5d39160f89201d1929ee25558", "output_hash": "3781b0f7175e64bec5fc4e06e8c55b65284738b62891371384a1ee1dc8c3cfe9"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "41e28ab69e7cc1f9d0c1c5e2", "attempt_id": "0c0935fcda86", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:18:38.174809Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 6194, "message_tokens": 6098, "message_utf8_bytes": 30224, "cache_tensor_bytes": null, "serialized_payload_bytes": 30224, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.29495800845324993, "ttft_ms": 13.988291029818356, "decode_ms": 96.41641599591821, "receiver_prefill_ms": 723.602292011492}, "total_wall_ms": 843.8504169462249, "ttft_ms": 13.988291029818356, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Lithmunpra-16", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S1"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "43187207d8fddf248ff8736f810b18469b7564f5d39160f89201d1929ee25558", "output_hash": "26fc169ceaba84d7f292defd76062949b1b87a9a87be89b5c857b8642cc75061"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "8574f9671a83f59c449a6ab2", "attempt_id": "1fbdfd89a1c9", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:18:39.035095Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 6187, "message_tokens": 6098, "message_utf8_bytes": 30224, "cache_tensor_bytes": null, "serialized_payload_bytes": 30224, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.29495800845324993, "ttft_ms": 13.886874890886247, "decode_ms": 110.98637501709163, "receiver_prefill_ms": 725.8460000157356}, "total_wall_ms": 860.6464159674942, "ttft_ms": 13.886874890886247, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "24485.45", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S1"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "43187207d8fddf248ff8736f810b18469b7564f5d39160f89201d1929ee25558", "output_hash": "7536bd6ac9283595bb370b563b72ea189d5a9c9e0d51df8a55df58c805878d5f"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "639fbdd030a0b71e52fa7bc9", "attempt_id": "96fd27ea22fb", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:18:39.915004Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 6188, "message_tokens": 6098, "message_utf8_bytes": 30224, "cache_tensor_bytes": null, "serialized_payload_bytes": 30224, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.29495800845324993, "ttft_ms": 14.10274999216199, "decode_ms": 13.599374913610518, "receiver_prefill_ms": 727.9752079630271}, "total_wall_ms": 765.514916041866, "ttft_ms": 14.10274999216199, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S1"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "43187207d8fddf248ff8736f810b18469b7564f5d39160f89201d1929ee25558", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "1fd5a4857609398a1102370d", "attempt_id": "d098f7b47762", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:18:40.697137Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 6189, "message_tokens": 6098, "message_utf8_bytes": 30224, "cache_tensor_bytes": null, "serialized_payload_bytes": 30224, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.29495800845324993, "ttft_ms": 13.7310839490965, "decode_ms": 14.492499991320074, "receiver_prefill_ms": 732.9031249973923}, "total_wall_ms": 770.9465409861878, "ttft_ms": 13.7310839490965, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S1"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "43187207d8fddf248ff8736f810b18469b7564f5d39160f89201d1929ee25558", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "421ca657c17f46a481343253", "attempt_id": "f91bc2755b69", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:18:56.053355Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 1125, "message_tokens": 1037, "message_utf8_bytes": 3019, "cache_tensor_bytes": null, "serialized_payload_bytes": 3019, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14568.443833035417, "ttft_ms": 13.68470792658627, "decode_ms": 94.6331248851493, "receiver_prefill_ms": 121.75408296752721}, "total_wall_ms": 14800.687498995103, "ttft_ms": 13.68470792658627, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7755450", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S1"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "923643230b058a72aadb3a5f7800b1b6daa55c0aa8c59e40f6f503c4162f37ae", "output_hash": "a30221bc7c855d26c2bda21a0a21fda10c953a05d985cec4e1dfa5a6a372d5d3"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "63bdc215f729458570b7fa21", "attempt_id": "8785b5c3c642", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:18:56.315650Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 1127, "message_tokens": 1037, "message_utf8_bytes": 3019, "cache_tensor_bytes": null, "serialized_payload_bytes": 3019, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14568.443833035417, "ttft_ms": 13.413708074949682, "decode_ms": 136.75029098521918, "receiver_prefill_ms": 121.77279102616012}, "total_wall_ms": 14842.37495798152, "ttft_ms": 13.413708074949682, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "RY-99D-9177", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S1"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "923643230b058a72aadb3a5f7800b1b6daa55c0aa8c59e40f6f503c4162f37ae", "output_hash": "abc5e1adc6533afa7ee1683a2195252eb090b08b52392c8150ccf6a387b14df6"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "0e49da9b6c0746818237940b", "attempt_id": "c5773200672e", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:18:56.607050Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 1133, "message_tokens": 1037, "message_utf8_bytes": 3019, "cache_tensor_bytes": null, "serialized_payload_bytes": 3019, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14568.443833035417, "ttft_ms": 13.3594999788329, "decode_ms": 94.7869170922786, "receiver_prefill_ms": 121.79975002072752}, "total_wall_ms": 14800.363708054647, "ttft_ms": 13.3594999788329, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7430110", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S1"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "923643230b058a72aadb3a5f7800b1b6daa55c0aa8c59e40f6f503c4162f37ae", "output_hash": "3781b0f7175e64bec5fc4e06e8c55b65284738b62891371384a1ee1dc8c3cfe9"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "4da34b4b4fd14819324d6da0", "attempt_id": "35355b06892f", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:18:56.856237Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 1126, "message_tokens": 1037, "message_utf8_bytes": 3019, "cache_tensor_bytes": null, "serialized_payload_bytes": 3019, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14568.443833035417, "ttft_ms": 13.414583052508533, "decode_ms": 95.17995896749198, "receiver_prefill_ms": 121.3852089131251}, "total_wall_ms": 14800.350458011962, "ttft_ms": 13.414583052508533, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "9844780", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S1"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "923643230b058a72aadb3a5f7800b1b6daa55c0aa8c59e40f6f503c4162f37ae", "output_hash": "771079f9bfa4045cea40b41caf04207b1c83e8ae61d7b82ef1421eae53276f16"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "998ae22dfbbf415c4db031e3", "attempt_id": "27ec5cb603ad", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:18:57.104620Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 1135, "message_tokens": 1037, "message_utf8_bytes": 3019, "cache_tensor_bytes": null, "serialized_payload_bytes": 3019, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14568.443833035417, "ttft_ms": 13.524916023015976, "decode_ms": 94.70287500880659, "receiver_prefill_ms": 121.68145796749741}, "total_wall_ms": 14800.330500118434, "ttft_ms": 13.524916023015976, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Lithmunpra-16", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S1"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "923643230b058a72aadb3a5f7800b1b6daa55c0aa8c59e40f6f503c4162f37ae", "output_hash": "26fc169ceaba84d7f292defd76062949b1b87a9a87be89b5c857b8642cc75061"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c35822455e8d4e4c50e88d56", "attempt_id": "06fd506aea99", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:18:57.353575Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 1128, "message_tokens": 1037, "message_utf8_bytes": 3019, "cache_tensor_bytes": null, "serialized_payload_bytes": 3019, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14568.443833035417, "ttft_ms": 13.405374949797988, "decode_ms": 94.17425002902746, "receiver_prefill_ms": 121.06724991463125}, "total_wall_ms": 14799.042582977563, "ttft_ms": 13.405374949797988, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "2349149", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S1"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "923643230b058a72aadb3a5f7800b1b6daa55c0aa8c59e40f6f503c4162f37ae", "output_hash": "4eeca6d8e49a7970739a5f11e5ee72c7e4ffa82281a6107cd1aba7577f873c2e"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "8964ff16b24fc9a9e056dcd3", "attempt_id": "fbb88df59f1d", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:18:57.600909Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 1129, "message_tokens": 1037, "message_utf8_bytes": 3019, "cache_tensor_bytes": null, "serialized_payload_bytes": 3019, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14568.443833035417, "ttft_ms": 13.800707994960248, "decode_ms": 13.514500111341476, "receiver_prefill_ms": 121.16412492468953}, "total_wall_ms": 14718.836083076894, "ttft_ms": 13.800707994960248, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S1"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "923643230b058a72aadb3a5f7800b1b6daa55c0aa8c59e40f6f503c4162f37ae", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "85e7324ec9df99002723d012", "attempt_id": "7a9bba839190", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:18:57.768211Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 1130, "message_tokens": 1037, "message_utf8_bytes": 3019, "cache_tensor_bytes": null, "serialized_payload_bytes": 3019, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14568.443833035417, "ttft_ms": 13.47183296456933, "decode_ms": 13.410374987870455, "receiver_prefill_ms": 120.74191705323756}, "total_wall_ms": 14718.03487406578, "ttft_ms": 13.47183296456933, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S1"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "923643230b058a72aadb3a5f7800b1b6daa55c0aa8c59e40f6f503c4162f37ae", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "46f541c99635ce6ee74b9dd5", "attempt_id": "62ed257961de", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:19:11.597743Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 16}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 732, "message_tokens": 641, "message_utf8_bytes": 1781, "cache_tensor_bytes": null, "serialized_payload_bytes": 1781, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13660.680916975252, "ttft_ms": 13.472417020238936, "decode_ms": 122.21016699913889, "receiver_prefill_ms": 78.930624993518}, "total_wall_ms": 13876.875126035884, "ttft_ms": 13.472417020238936, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7,755,450", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S1"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "365487d3000992b5a4f4f8607767cbfe611d2005902079f94756241a57c3c75a", "output_hash": "50ce1bf5a6a1535bae98114196498afa876eadf9e5680cdc8087d8bc8f71357e"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "ea723ab4e644c5479b7a53c5", "attempt_id": "87cb63fea6fe", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:19:11.831704Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 16}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 734, "message_tokens": 641, "message_utf8_bytes": 1781, "cache_tensor_bytes": null, "serialized_payload_bytes": 1781, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13660.680916975252, "ttft_ms": 13.688125065527856, "decode_ms": 13.649791944772005, "receiver_prefill_ms": 78.54274997953326}, "total_wall_ms": 13768.053208943456, "ttft_ms": 13.688125065527856, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S1"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "365487d3000992b5a4f4f8607767cbfe611d2005902079f94756241a57c3c75a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c5274d8df9b49df9305a154f", "attempt_id": "c2253c50f523", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:19:11.957021Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 16}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 740, "message_tokens": 641, "message_utf8_bytes": 1781, "cache_tensor_bytes": null, "serialized_payload_bytes": 1781, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13660.680916975252, "ttft_ms": 13.648124993778765, "decode_ms": 97.22825000062585, "receiver_prefill_ms": 81.654290901497}, "total_wall_ms": 13854.71945896279, "ttft_ms": 13.648124993778765, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7430110", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S1"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "365487d3000992b5a4f4f8607767cbfe611d2005902079f94756241a57c3c75a", "output_hash": "3781b0f7175e64bec5fc4e06e8c55b65284738b62891371384a1ee1dc8c3cfe9"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "9a14429a143d00a6d4c25936", "attempt_id": "4d7622f8b03a", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:19:12.168399Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 16}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 733, "message_tokens": 641, "message_utf8_bytes": 1781, "cache_tensor_bytes": null, "serialized_payload_bytes": 1781, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13660.680916975252, "ttft_ms": 13.81620904430747, "decode_ms": 95.65229201689363, "receiver_prefill_ms": 78.22095800656825}, "total_wall_ms": 13849.84495898243, "ttft_ms": 13.81620904430747, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7430110", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S1"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "365487d3000992b5a4f4f8607767cbfe611d2005902079f94756241a57c3c75a", "output_hash": "3781b0f7175e64bec5fc4e06e8c55b65284738b62891371384a1ee1dc8c3cfe9"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "e18195b2538248c2eeef8a91", "attempt_id": "2b93984d1837", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:19:12.374648Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 16}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 742, "message_tokens": 641, "message_utf8_bytes": 1781, "cache_tensor_bytes": null, "serialized_payload_bytes": 1781, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13660.680916975252, "ttft_ms": 13.577874982729554, "decode_ms": 94.65162502601743, "receiver_prefill_ms": 81.0630830237642}, "total_wall_ms": 13851.478958968073, "ttft_ms": 13.577874982729554, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Lithmunpra-16", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S1"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "365487d3000992b5a4f4f8607767cbfe611d2005902079f94756241a57c3c75a", "output_hash": "26fc169ceaba84d7f292defd76062949b1b87a9a87be89b5c857b8642cc75061"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "ce623b8775035661108d42f4", "attempt_id": "183618bdca9d", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:19:12.582838Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 16}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 735, "message_tokens": 641, "message_utf8_bytes": 1781, "cache_tensor_bytes": null, "serialized_payload_bytes": 1781, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13660.680916975252, "ttft_ms": 13.279416016303003, "decode_ms": 135.26512507814914, "receiver_prefill_ms": 78.2376250717789}, "total_wall_ms": 13888.924126047641, "ttft_ms": 13.279416016303003, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "23,978,337", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S1"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "365487d3000992b5a4f4f8607767cbfe611d2005902079f94756241a57c3c75a", "output_hash": "d809cb2d0d2d9b0f5de91d048c23f95338b4bd65b04c88289493f00ab52a0d50"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "2b84e43ec568e621b9dbc308", "attempt_id": "7f49f1f54f88", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:19:12.828519Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 16}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 736, "message_tokens": 641, "message_utf8_bytes": 1781, "cache_tensor_bytes": null, "serialized_payload_bytes": 1781, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13660.680916975252, "ttft_ms": 13.422542018815875, "decode_ms": 13.338292017579079, "receiver_prefill_ms": 78.25758401304483}, "total_wall_ms": 13767.2751249047, "ttft_ms": 13.422542018815875, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S1"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "365487d3000992b5a4f4f8607767cbfe611d2005902079f94756241a57c3c75a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "d66442109f89d26df6b541fd", "attempt_id": "46094828b386", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:19:12.952836Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 16}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 737, "message_tokens": 641, "message_utf8_bytes": 1781, "cache_tensor_bytes": null, "serialized_payload_bytes": 1781, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13660.680916975252, "ttft_ms": 13.576583005487919, "decode_ms": 13.33620899822563, "receiver_prefill_ms": 78.76191707327962}, "total_wall_ms": 13767.90945790708, "ttft_ms": 13.576583005487919, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S1"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "365487d3000992b5a4f4f8607767cbfe611d2005902079f94756241a57c3c75a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a270348d351c25de7e5c588b", "attempt_id": "60a6b002258e", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:19:26.767826Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 753, "message_tokens": 663, "message_utf8_bytes": 1743, "cache_tensor_bytes": null, "serialized_payload_bytes": 1743, "side_information_bytes": 0, "fetched_bytes": 931, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13689.993625041097, "ttft_ms": 13.337207958102226, "decode_ms": 13.5240419767797, "receiver_prefill_ms": 81.3775829738006}, "total_wall_ms": 13802.4187090341, "ttft_ms": 13.337207958102226, "input_usage": {"lookups": 3}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S1"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "add7d7753aa3276c4815b25d676461fda8e169670ef60f11fa373afb875c4b47", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "5d7fcfffc789deed510dd22b", "attempt_id": "183bd623a2d9", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:19:26.897724Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 702, "message_tokens": 610, "message_utf8_bytes": 1619, "cache_tensor_bytes": null, "serialized_payload_bytes": 1619, "side_information_bytes": 0, "fetched_bytes": 822, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13689.993625041097, "ttft_ms": 13.668542029336095, "decode_ms": 106.87637492083013, "receiver_prefill_ms": 76.95299992337823}, "total_wall_ms": 13889.35858302284, "ttft_ms": 13.668542029336095, "input_usage": {"lookups": 3}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "RY-99D-91", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S1"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "50f0f190516ef199a91b9e1e132ced9ead362406fb0696f115470ebeabe71dd5", "output_hash": "0d4fc4444a5340bd4c2ae14e40ac36425ee1abc465c137c436e12006bb11e156"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "f4dc3eb4ae7ac6425479f80d", "attempt_id": "093788719a55", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:19:27.115151Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 752, "message_tokens": 654, "message_utf8_bytes": 1712, "cache_tensor_bytes": null, "serialized_payload_bytes": 1712, "side_information_bytes": 0, "fetched_bytes": 897, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13689.993625041097, "ttft_ms": 13.315209071151912, "decode_ms": 93.63316697999835, "receiver_prefill_ms": 81.06729201972485}, "total_wall_ms": 13879.526959033683, "ttft_ms": 13.315209071151912, "input_usage": {"lookups": 3}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7430110", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S1"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "07a2ab299fffdece8aada5829983283a9cf3a46fb4d8824bb7a0db4950647aa5", "output_hash": "3781b0f7175e64bec5fc4e06e8c55b65284738b62891371384a1ee1dc8c3cfe9"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "1596eb658cd3308bf6be0b14", "attempt_id": "358a1823c0a8", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:19:27.322492Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 745, "message_tokens": 654, "message_utf8_bytes": 1712, "cache_tensor_bytes": null, "serialized_payload_bytes": 1712, "side_information_bytes": 0, "fetched_bytes": 897, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13689.993625041097, "ttft_ms": 13.318665907718241, "decode_ms": 93.87012501247227, "receiver_prefill_ms": 81.0636670794338}, "total_wall_ms": 13879.734875052236, "ttft_ms": 13.318665907718241, "input_usage": {"lookups": 3}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "9844780", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S1"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "07a2ab299fffdece8aada5829983283a9cf3a46fb4d8824bb7a0db4950647aa5", "output_hash": "771079f9bfa4045cea40b41caf04207b1c83e8ae61d7b82ef1421eae53276f16"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a39d37a86c21ff025c29274a", "attempt_id": "a368e7c702a8", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:19:27.529608Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 718, "message_tokens": 618, "message_utf8_bytes": 1642, "cache_tensor_bytes": null, "serialized_payload_bytes": 1642, "side_information_bytes": 0, "fetched_bytes": 833, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13689.993625041097, "ttft_ms": 13.58050003182143, "decode_ms": 13.500374974682927, "receiver_prefill_ms": 78.01449997350574}, "total_wall_ms": 13796.494708978571, "ttft_ms": 13.58050003182143, "input_usage": {"lookups": 3}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S1"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "21b7ed9d232bb4fdb17e63edc0732e829ddc65f3e5c5a6c5a26d56a8d0473245", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "bc42046a66492217aded9c3c", "attempt_id": "5928bd9e446e", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:19:27.653819Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 756, "message_tokens": 663, "message_utf8_bytes": 1743, "cache_tensor_bytes": null, "serialized_payload_bytes": 1743, "side_information_bytes": 0, "fetched_bytes": 931, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13689.993625041097, "ttft_ms": 13.326750020496547, "decode_ms": 133.90470808371902, "receiver_prefill_ms": 81.66949998121709}, "total_wall_ms": 13920.343625009991, "ttft_ms": 13.326750020496547, "input_usage": {"lookups": 3}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "17,292,865", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S1"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "add7d7753aa3276c4815b25d676461fda8e169670ef60f11fa373afb875c4b47", "output_hash": "41c4eae1e47c07669e6d4bc29c7ad346ec84c068b47eaad8b45213d89e7d7382"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "7d2fa125a2c55d480e7b6a83", "attempt_id": "3502f0b90fe2", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:19:27.900951Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 755, "message_tokens": 661, "message_utf8_bytes": 1700, "cache_tensor_bytes": null, "serialized_payload_bytes": 1700, "side_information_bytes": 0, "fetched_bytes": 891, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13689.993625041097, "ttft_ms": 13.570124981924891, "decode_ms": 13.361333985812962, "receiver_prefill_ms": 80.93329204712063}, "total_wall_ms": 13799.41175004933, "ttft_ms": 13.570124981924891, "input_usage": {"lookups": 3}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S1"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "bda4a5e46902b86f91bca852b00ec94fe6464ef2bc54517be64b81893a49cfbb", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "1a417ce94bc9e8805a5bc260", "attempt_id": "ee85d1eee551", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:19:28.027392Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 756, "message_tokens": 661, "message_utf8_bytes": 1700, "cache_tensor_bytes": null, "serialized_payload_bytes": 1700, "side_information_bytes": 0, "fetched_bytes": 891, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13689.993625041097, "ttft_ms": 13.428292004391551, "decode_ms": 13.613875024020672, "receiver_prefill_ms": 80.45008289627731}, "total_wall_ms": 13798.984166001901, "ttft_ms": 13.428292004391551, "input_usage": {"lookups": 3}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S1"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "bda4a5e46902b86f91bca852b00ec94fe6464ef2bc54517be64b81893a49cfbb", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "9fabb9ca22f82530fe9c6fb0", "attempt_id": "3ca67471b566", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:19:28.901166Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 227672064}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 44, "message_tokens": null, "message_utf8_bytes": 30198, "cache_tensor_bytes": 227672064, "serialized_payload_bytes": 227672064, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 747.5203330395743, "ttft_ms": 53.1677920371294, "decode_ms": 106.65008297655731, "receiver_prefill_ms": 0.0}, "total_wall_ms": 916.9128750218078, "ttft_ms": 53.1677920371294, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4076350", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S1"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "a29b3a497fa3db91ff78701e0096c2d007f43888aabd09e84e1cbd96588a8f94"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "eef651cef16461c0fd5dd239", "attempt_id": "4ed772a8e1b5", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:19:29.087717Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 227672064}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 46, "message_tokens": null, "message_utf8_bytes": 30198, "cache_tensor_bytes": 227672064, "serialized_payload_bytes": 227672064, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 747.5203330395743, "ttft_ms": 47.33983299229294, "decode_ms": 150.1969579840079, "receiver_prefill_ms": 0.0}, "total_wall_ms": 952.2020410513505, "ttft_ms": 47.33983299229294, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "RY-99D-9177", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S1"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "abc5e1adc6533afa7ee1683a2195252eb090b08b52392c8150ccf6a387b14df6"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "0b12ea7d41af4e990858b418", "attempt_id": "9046ef9df298", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:19:29.310256Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 227672064}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 52, "message_tokens": null, "message_utf8_bytes": 30198, "cache_tensor_bytes": 227672064, "serialized_payload_bytes": 227672064, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 747.5203330395743, "ttft_ms": 45.9299999056384, "decode_ms": 105.06845801137388, "receiver_prefill_ms": 0.0}, "total_wall_ms": 906.0746249742806, "ttft_ms": 45.9299999056384, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7430110", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S1"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "3781b0f7175e64bec5fc4e06e8c55b65284738b62891371384a1ee1dc8c3cfe9"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "4b6a2e7ff7dbea00017ee27e", "attempt_id": "661bfb3b54f4", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:19:29.486470Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 227672064}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 45, "message_tokens": null, "message_utf8_bytes": 30198, "cache_tensor_bytes": 227672064, "serialized_payload_bytes": 227672064, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 747.5203330395743, "ttft_ms": 48.842750024050474, "decode_ms": 135.50125004258007, "receiver_prefill_ms": 0.0}, "total_wall_ms": 938.460958073847, "ttft_ms": 48.842750024050474, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7,430,110", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S1"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "6ec8f3686ca12e2212d5f0b03da40f6c90866d495314345bafd9be7b9638f772"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "fd14286d07885f23d746b9b1", "attempt_id": "3acc5e5aa387", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:19:29.695072Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 227672064}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 54, "message_tokens": null, "message_utf8_bytes": 30198, "cache_tensor_bytes": 227672064, "serialized_payload_bytes": 227672064, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 747.5203330395743, "ttft_ms": 47.50854196026921, "decode_ms": 105.6153749814257, "receiver_prefill_ms": 0.0}, "total_wall_ms": 908.3085830789059, "ttft_ms": 47.50854196026921, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Lithmunpra-16", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S1"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "26fc169ceaba84d7f292defd76062949b1b87a9a87be89b5c857b8642cc75061"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "7465eb4532dc499658252907", "attempt_id": "05145d2fc971", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:19:29.873558Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 227672064}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 30198, "cache_tensor_bytes": 227672064, "serialized_payload_bytes": 227672064, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 747.5203330395743, "ttft_ms": 48.1460839509964, "decode_ms": 120.02387503162026, "receiver_prefill_ms": 0.0}, "total_wall_ms": 923.1482499744743, "ttft_ms": 48.1460839509964, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "25452.86", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S1"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "e4f795cfea3fc8dee9e60160ca4059eb40b14672d0087d490aaaa401a8a37869"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "2bfda3b952b0fd44d6d81b4d", "attempt_id": "437adb8c716d", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:19:30.066925Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 227672064}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 48, "message_tokens": null, "message_utf8_bytes": 30198, "cache_tensor_bytes": 227672064, "serialized_payload_bytes": 227672064, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 747.5203330395743, "ttft_ms": 45.564542058855295, "decode_ms": 15.175167005509138, "receiver_prefill_ms": 0.0}, "total_wall_ms": 815.865291049704, "ttft_ms": 45.564542058855295, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S1"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "2d93f3d87952deb8d4ebb628", "attempt_id": "f5c91130cf7a", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:19:30.152640Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 227672064}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 49, "message_tokens": null, "message_utf8_bytes": 30198, "cache_tensor_bytes": 227672064, "serialized_payload_bytes": 227672064, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 747.5203330395743, "ttft_ms": 48.15216606948525, "decode_ms": 15.102791949175298, "receiver_prefill_ms": 0.0}, "total_wall_ms": 818.2527490425855, "ttft_ms": 48.15216606948525, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S1"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "7c796f83749e3c7334e86485", "attempt_id": "fda935a8ac76", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:19:30.241010Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 6298, "message_tokens": 6212, "message_utf8_bytes": 30728, "cache_tensor_bytes": null, "serialized_payload_bytes": 30728, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.029916991479694843, "ttft_ms": 14.252708991989493, "decode_ms": 98.89424999710172, "receiver_prefill_ms": 753.2849169801921}, "total_wall_ms": 877.3673330433667, "ttft_ms": 14.252708991989493, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7755450", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S3"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a000885caf496d5c7458e2004b2ba315e216792aa6e8ef1811d22473294c8ccc", "output_hash": "a30221bc7c855d26c2bda21a0a21fda10c953a05d985cec4e1dfa5a6a372d5d3"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "8ecc4979df4585fb9921b163", "attempt_id": "ecb0e5537fc8", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:19:31.135931Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 6300, "message_tokens": 6212, "message_utf8_bytes": 30728, "cache_tensor_bytes": null, "serialized_payload_bytes": 30728, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.029916991479694843, "ttft_ms": 13.994709006510675, "decode_ms": 138.72724992688745, "receiver_prefill_ms": 769.0287089208141}, "total_wall_ms": 931.6374170593917, "ttft_ms": 13.994709006510675, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "RY-99D-9177", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S3"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a000885caf496d5c7458e2004b2ba315e216792aa6e8ef1811d22473294c8ccc", "output_hash": "abc5e1adc6533afa7ee1683a2195252eb090b08b52392c8150ccf6a387b14df6"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "179c0ed0781c23b10f99275d", "attempt_id": "035c87e9c5b8", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:19:32.084907Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 6306, "message_tokens": 6212, "message_utf8_bytes": 30728, "cache_tensor_bytes": null, "serialized_payload_bytes": 30728, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.029916991479694843, "ttft_ms": 13.965707970783114, "decode_ms": 98.00754091702402, "receiver_prefill_ms": 782.5275840004906}, "total_wall_ms": 904.2063750093803, "ttft_ms": 13.965707970783114, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7430110", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S3"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a000885caf496d5c7458e2004b2ba315e216792aa6e8ef1811d22473294c8ccc", "output_hash": "3781b0f7175e64bec5fc4e06e8c55b65284738b62891371384a1ee1dc8c3cfe9"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "37dd020e00f72d15a1a1c050", "attempt_id": "003c6eff1398", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:19:33.005896Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 6299, "message_tokens": 6212, "message_utf8_bytes": 30728, "cache_tensor_bytes": null, "serialized_payload_bytes": 30728, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.029916991479694843, "ttft_ms": 14.370624907314777, "decode_ms": 97.09924994967878, "receiver_prefill_ms": 806.5085420385003}, "total_wall_ms": 927.7734169736505, "ttft_ms": 14.370624907314777, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7430110", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S3"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a000885caf496d5c7458e2004b2ba315e216792aa6e8ef1811d22473294c8ccc", "output_hash": "3781b0f7175e64bec5fc4e06e8c55b65284738b62891371384a1ee1dc8c3cfe9"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "3a79d45d519598e970b5bc55", "attempt_id": "db558a5761dd", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:19:33.951382Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 6308, "message_tokens": 6212, "message_utf8_bytes": 30728, "cache_tensor_bytes": null, "serialized_payload_bytes": 30728, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.029916991479694843, "ttft_ms": 13.975166017189622, "decode_ms": 96.94312501233071, "receiver_prefill_ms": 826.1763750342652}, "total_wall_ms": 946.8585419235751, "ttft_ms": 13.975166017189622, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Lithmunpra-16", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S3"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a000885caf496d5c7458e2004b2ba315e216792aa6e8ef1811d22473294c8ccc", "output_hash": "26fc169ceaba84d7f292defd76062949b1b87a9a87be89b5c857b8642cc75061"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "6501752a62207d308daf5ba1", "attempt_id": "f4f22acb86df", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:19:34.915879Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 6301, "message_tokens": 6212, "message_utf8_bytes": 30728, "cache_tensor_bytes": null, "serialized_payload_bytes": 30728, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.029916991479694843, "ttft_ms": 14.20145807787776, "decode_ms": 97.6001659873873, "receiver_prefill_ms": 834.6187500283122}, "total_wall_ms": 956.2408759957179, "ttft_ms": 14.20145807787776, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "2474456", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S3"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a000885caf496d5c7458e2004b2ba315e216792aa6e8ef1811d22473294c8ccc", "output_hash": "5ca93f7b1d6281b825e236dcc18e7b221bcaf8fb291cc10db1e38cbb25d200b0"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "906ce71b0606a27757c05a3e", "attempt_id": "1c1e77c19f42", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:19:35.890134Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 6302, "message_tokens": 6212, "message_utf8_bytes": 30728, "cache_tensor_bytes": null, "serialized_payload_bytes": 30728, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.029916991479694843, "ttft_ms": 14.15020797867328, "decode_ms": 13.982250005938113, "receiver_prefill_ms": 849.564041942358}, "total_wall_ms": 887.4901259550825, "ttft_ms": 14.15020797867328, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S3"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a000885caf496d5c7458e2004b2ba315e216792aa6e8ef1811d22473294c8ccc", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "5ca9873ed9e56bb15113237f", "attempt_id": "153d3c00cca9", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:19:36.794879Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 6303, "message_tokens": 6212, "message_utf8_bytes": 30728, "cache_tensor_bytes": null, "serialized_payload_bytes": 30728, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.029916991479694843, "ttft_ms": 14.300167094916105, "decode_ms": 13.983041048049927, "receiver_prefill_ms": 822.3296659998596}, "total_wall_ms": 860.465083969757, "ttft_ms": 14.300167094916105, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S3"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a000885caf496d5c7458e2004b2ba315e216792aa6e8ef1811d22473294c8ccc", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "11091a7a09d68b4e076181bb", "attempt_id": "9205307ea457", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:20:03.921991Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1864}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 1977, "message_tokens": 1889, "message_utf8_bytes": 6624, "cache_tensor_bytes": null, "serialized_payload_bytes": 6624, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26248.771040933207, "ttft_ms": 13.764750096015632, "decode_ms": 97.55845798645169, "receiver_prefill_ms": 208.10325001366436}, "total_wall_ms": 26576.01362396963, "ttft_ms": 13.764750096015632, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7755450", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S3"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "86068d22548bd2159508fa4ff5d79336a67485e7675e04f9808b27b46c477f5f", "output_hash": "a30221bc7c855d26c2bda21a0a21fda10c953a05d985cec4e1dfa5a6a372d5d3"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "80263381b1cc9a3866cfff48", "attempt_id": "9f48b9491014", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:20:04.268460Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1864}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 1979, "message_tokens": 1889, "message_utf8_bytes": 6624, "cache_tensor_bytes": null, "serialized_payload_bytes": 6624, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26248.771040933207, "ttft_ms": 13.592166942544281, "decode_ms": 136.2872909521684, "receiver_prefill_ms": 206.71658299397677}, "total_wall_ms": 26608.499790891074, "ttft_ms": 13.592166942544281, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "RY-99D-9177", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S3"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "86068d22548bd2159508fa4ff5d79336a67485e7675e04f9808b27b46c477f5f", "output_hash": "abc5e1adc6533afa7ee1683a2195252eb090b08b52392c8150ccf6a387b14df6"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b9d2e2a0fd75be14e86536fd", "attempt_id": "d40f76840e75", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:20:04.645324Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1864}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 1985, "message_tokens": 1889, "message_utf8_bytes": 6624, "cache_tensor_bytes": null, "serialized_payload_bytes": 6624, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26248.771040933207, "ttft_ms": 13.791040983051062, "decode_ms": 96.58833406865597, "receiver_prefill_ms": 256.19116600137204}, "total_wall_ms": 26618.5372909531, "ttft_ms": 13.791040983051062, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7430110", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S3"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "86068d22548bd2159508fa4ff5d79336a67485e7675e04f9808b27b46c477f5f", "output_hash": "3781b0f7175e64bec5fc4e06e8c55b65284738b62891371384a1ee1dc8c3cfe9"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "f42cda91dfe85af2bd325cba", "attempt_id": "4af82b4ab1cf", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:20:05.033706Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1864}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 1978, "message_tokens": 1889, "message_utf8_bytes": 6624, "cache_tensor_bytes": null, "serialized_payload_bytes": 6624, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26248.771040933207, "ttft_ms": 13.669000007212162, "decode_ms": 94.98058306053281, "receiver_prefill_ms": 234.55924994777888}, "total_wall_ms": 26595.19416594412, "ttft_ms": 13.669000007212162, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "9844780", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S3"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "86068d22548bd2159508fa4ff5d79336a67485e7675e04f9808b27b46c477f5f", "output_hash": "771079f9bfa4045cea40b41caf04207b1c83e8ae61d7b82ef1421eae53276f16"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "86e6a8b62ee54bc60985507e", "attempt_id": "7e08f9271333", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:20:05.399820Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1864}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 1987, "message_tokens": 1889, "message_utf8_bytes": 6624, "cache_tensor_bytes": null, "serialized_payload_bytes": 6624, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26248.771040933207, "ttft_ms": 13.704707962460816, "decode_ms": 68.89912497717887, "receiver_prefill_ms": 230.66179198212922}, "total_wall_ms": 26565.226956969127, "ttft_ms": 13.704707962460816, "input_usage": {}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Soltek-12", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S3"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "86068d22548bd2159508fa4ff5d79336a67485e7675e04f9808b27b46c477f5f", "output_hash": "afeabbb13c97073235bda8ee98e604768e20ddd816bcc53f1cf102c9aefab62c"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "429003cb4fcab969800fee09", "attempt_id": "5c3d38ab35bf", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:20:05.733928Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1864}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 1980, "message_tokens": 1889, "message_utf8_bytes": 6624, "cache_tensor_bytes": null, "serialized_payload_bytes": 6624, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26248.771040933207, "ttft_ms": 13.960084063000977, "decode_ms": 108.09291701298207, "receiver_prefill_ms": 231.92087491042912}, "total_wall_ms": 26606.3520408934, "ttft_ms": 13.960084063000977, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "20895456", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S3"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "86068d22548bd2159508fa4ff5d79336a67485e7675e04f9808b27b46c477f5f", "output_hash": "dadcd418968cd77fef94ff7583937d817d2fb99194a7bfcc17d0ce1b03b7cb0c"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "215c7e3f49059c79248ee16c", "attempt_id": "d8749e330142", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:20:06.110143Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1864}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 1981, "message_tokens": 1889, "message_utf8_bytes": 6624, "cache_tensor_bytes": null, "serialized_payload_bytes": 6624, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26248.771040933207, "ttft_ms": 13.648665975779295, "decode_ms": 13.738458044826984, "receiver_prefill_ms": 228.70824998244643}, "total_wall_ms": 26507.961333030835, "ttft_ms": 13.648665975779295, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S3"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "86068d22548bd2159508fa4ff5d79336a67485e7675e04f9808b27b46c477f5f", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "545d011c427719f761d667d9", "attempt_id": "e973726b14fe", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:20:06.387790Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1864}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 1982, "message_tokens": 1889, "message_utf8_bytes": 6624, "cache_tensor_bytes": null, "serialized_payload_bytes": 6624, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26248.771040933207, "ttft_ms": 14.104582951404154, "decode_ms": 13.540999963879585, "receiver_prefill_ms": 498.6093749757856}, "total_wall_ms": 26778.084374847822, "ttft_ms": 14.104582951404154, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S3"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "86068d22548bd2159508fa4ff5d79336a67485e7675e04f9808b27b46c477f5f", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "d1887201fff650b7e0d69f05", "attempt_id": "384034593513", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:20:33.119420Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 27}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 1148, "message_tokens": 1057, "message_utf8_bytes": 2842, "cache_tensor_bytes": null, "serialized_payload_bytes": 2842, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26184.120082994923, "ttft_ms": 14.312042039819062, "decode_ms": 125.3163330256939, "receiver_prefill_ms": 727.5804160162807}, "total_wall_ms": 27053.318958031014, "ttft_ms": 14.312042039819062, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4,076,350", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S3"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "52b89886615f4d609df571409a1ad65685797c447821f343dc5cfe0223a476ab", "output_hash": "34656deee2371a2f5af82afa7ac5bc2813c79ea1116cda9cc6b884cb30096edb"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "d6eb0e0b4035dd0c540d7460", "attempt_id": "76885a75e748", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:20:34.012206Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 27}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 1150, "message_tokens": 1057, "message_utf8_bytes": 2842, "cache_tensor_bytes": null, "serialized_payload_bytes": 2842, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26184.120082994923, "ttft_ms": 14.01133299805224, "decode_ms": 139.04595898929983, "receiver_prefill_ms": 138.34470906294882}, "total_wall_ms": 26477.5160420686, "ttft_ms": 14.01133299805224, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "BG-94A-2920", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S3"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "52b89886615f4d609df571409a1ad65685797c447821f343dc5cfe0223a476ab", "output_hash": "d2a50219130b991d05fe1c9a88dcba36baed0c7f37529a15ad7582c47c756d67"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "ea120126e172cd59acb76961", "attempt_id": "507bf254b992", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:20:34.324369Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 27}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 1156, "message_tokens": 1057, "message_utf8_bytes": 2842, "cache_tensor_bytes": null, "serialized_payload_bytes": 2842, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26184.120082994923, "ttft_ms": 13.712332933209836, "decode_ms": 95.52933392114937, "receiver_prefill_ms": 125.33549999352545}, "total_wall_ms": 26420.506207970902, "ttft_ms": 13.712332933209836, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7430110", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S3"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "52b89886615f4d609df571409a1ad65685797c447821f343dc5cfe0223a476ab", "output_hash": "3781b0f7175e64bec5fc4e06e8c55b65284738b62891371384a1ee1dc8c3cfe9"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c1f6dd00856d2796b6daa4ba", "attempt_id": "90147a1c770d", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:20:34.578749Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 27}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 1149, "message_tokens": 1057, "message_utf8_bytes": 2842, "cache_tensor_bytes": null, "serialized_payload_bytes": 2842, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26184.120082994923, "ttft_ms": 14.0805410919711, "decode_ms": 98.3205420197919, "receiver_prefill_ms": 121.93445803131908}, "total_wall_ms": 26420.47533299774, "ttft_ms": 14.0805410919711, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7430110", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S3"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "52b89886615f4d609df571409a1ad65685797c447821f343dc5cfe0223a476ab", "output_hash": "3781b0f7175e64bec5fc4e06e8c55b65284738b62891371384a1ee1dc8c3cfe9"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "6c47ed72be9b2ae986f20b8e", "attempt_id": "ea3f8aa4a0a9", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:20:34.835238Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 27}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 1158, "message_tokens": 1057, "message_utf8_bytes": 2842, "cache_tensor_bytes": null, "serialized_payload_bytes": 2842, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26184.120082994923, "ttft_ms": 14.311500010080636, "decode_ms": 96.51587507687509, "receiver_prefill_ms": 125.29379199258983}, "total_wall_ms": 26422.27712494787, "ttft_ms": 14.311500010080636, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Lithmunpra-16", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S3"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "52b89886615f4d609df571409a1ad65685797c447821f343dc5cfe0223a476ab", "output_hash": "26fc169ceaba84d7f292defd76062949b1b87a9a87be89b5c857b8642cc75061"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "858f2585d87d0432f05105c0", "attempt_id": "fb87d3e75d81", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:20:35.094981Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 27}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 1151, "message_tokens": 1057, "message_utf8_bytes": 2842, "cache_tensor_bytes": null, "serialized_payload_bytes": 2842, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26184.120082994923, "ttft_ms": 13.822583016008139, "decode_ms": 96.05129202827811, "receiver_prefill_ms": 498.46916703972965}, "total_wall_ms": 26794.52879098244, "ttft_ms": 13.822583016008139, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "2620338", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S3"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "52b89886615f4d609df571409a1ad65685797c447821f343dc5cfe0223a476ab", "output_hash": "878fc5b932e75f219065be91ac0f38349f591fd38efaf6d3d53982d220f5c725"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "8e1a786c2a04d0ef5e4d33fa", "attempt_id": "658a9ed6ea8b", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:20:35.724368Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 27}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 1152, "message_tokens": 1057, "message_utf8_bytes": 2842, "cache_tensor_bytes": null, "serialized_payload_bytes": 2842, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26184.120082994923, "ttft_ms": 13.393832952715456, "decode_ms": 13.297167024575174, "receiver_prefill_ms": 121.63220904767513}, "total_wall_ms": 26334.28570802789, "ttft_ms": 13.393832952715456, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S3"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "52b89886615f4d609df571409a1ad65685797c447821f343dc5cfe0223a476ab", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "4c37d12e5ed198966c4d8d2f", "attempt_id": "f2d441acbeb3", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:20:35.892320Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 27}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 1153, "message_tokens": 1057, "message_utf8_bytes": 2842, "cache_tensor_bytes": null, "serialized_payload_bytes": 2842, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26184.120082994923, "ttft_ms": 14.003833988681436, "decode_ms": 13.976166024804115, "receiver_prefill_ms": 242.45466606225818}, "total_wall_ms": 26456.397749017924, "ttft_ms": 14.003833988681436, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S3"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "52b89886615f4d609df571409a1ad65685797c447821f343dc5cfe0223a476ab", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a098d240eb31d94a0740cbc8", "attempt_id": "172172b1d4a2", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:21:02.109030Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 2635, "message_tokens": 2545, "message_utf8_bytes": 6643, "cache_tensor_bytes": null, "serialized_payload_bytes": 6643, "side_information_bytes": 0, "fetched_bytes": 3548, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25926.604749984108, "ttft_ms": 14.239707961678505, "decode_ms": 97.55433304235339, "receiver_prefill_ms": 333.24425003957003}, "total_wall_ms": 26376.67970790062, "ttft_ms": 14.239707961678505, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7755450", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S3"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ea82bca55b3e0395c0cbaeda4a4f3918e461d53edcf7cf50f7f693eda7725b6d", "output_hash": "a30221bc7c855d26c2bda21a0a21fda10c953a05d985cec4e1dfa5a6a372d5d3"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "000a44acd0d19fe1cc0b89ae", "attempt_id": "c45ecca6549b", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:21:02.583077Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 2533, "message_tokens": 2441, "message_utf8_bytes": 6338, "cache_tensor_bytes": null, "serialized_payload_bytes": 6338, "side_information_bytes": 0, "fetched_bytes": 3245, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25926.604749984108, "ttft_ms": 13.492166064679623, "decode_ms": 135.7113749254495, "receiver_prefill_ms": 272.71483303047717}, "total_wall_ms": 26359.626333927736, "ttft_ms": 13.492166064679623, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "RY-99D-9177", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S3"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "c2a5f7190c4bfdc34be8bc736dc964d309c0bb88db412df6d8ce67d5f41cb6ef", "output_hash": "abc5e1adc6533afa7ee1683a2195252eb090b08b52392c8150ccf6a387b14df6"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "1e217fee6d28304a98c5bc51", "attempt_id": "d53ce619b97a", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:21:03.037832Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 2643, "message_tokens": 2545, "message_utf8_bytes": 6643, "cache_tensor_bytes": null, "serialized_payload_bytes": 6643, "side_information_bytes": 0, "fetched_bytes": 3548, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25926.604749984108, "ttft_ms": 13.71483295224607, "decode_ms": 96.19874996133149, "receiver_prefill_ms": 291.1176671041176}, "total_wall_ms": 26331.187707954086, "ttft_ms": 13.71483295224607, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7430110", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S3"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "c515914e4c8084b7a500fca1bc902b838177a6600be56f5530bcbab0833dc173", "output_hash": "3781b0f7175e64bec5fc4e06e8c55b65284738b62891371384a1ee1dc8c3cfe9"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "51a31c79149c4ffaf9ae7277", "attempt_id": "88a515189eb6", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:21:03.460106Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 2636, "message_tokens": 2545, "message_utf8_bytes": 6643, "cache_tensor_bytes": null, "serialized_payload_bytes": 6643, "side_information_bytes": 0, "fetched_bytes": 3548, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25926.604749984108, "ttft_ms": 14.141666935756803, "decode_ms": 96.29095799755305, "receiver_prefill_ms": 317.4919580342248}, "total_wall_ms": 26358.04241592996, "ttft_ms": 14.141666935756803, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "9844780", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S3"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "c515914e4c8084b7a500fca1bc902b838177a6600be56f5530bcbab0833dc173", "output_hash": "771079f9bfa4045cea40b41caf04207b1c83e8ae61d7b82ef1421eae53276f16"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "430d132fdd9f81ddc0a7545a", "attempt_id": "f6f8f3500ce7", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:21:03.912315Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 2597, "message_tokens": 2497, "message_utf8_bytes": 6418, "cache_tensor_bytes": null, "serialized_payload_bytes": 6418, "side_information_bytes": 0, "fetched_bytes": 3350, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25926.604749984108, "ttft_ms": 14.22812498640269, "decode_ms": 96.36258298996836, "receiver_prefill_ms": 331.90537500195205}, "total_wall_ms": 26373.54654201772, "ttft_ms": 14.22812498640269, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Lithmunpra-16", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S3"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "77600cbb0ef043ea0e4fd7276f9db7a25e7d5298c19159699a692be9208eb3c4", "output_hash": "26fc169ceaba84d7f292defd76062949b1b87a9a87be89b5c857b8642cc75061"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "39399f5d4d2529aa4e7901bc", "attempt_id": "9f99b0367ed4", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:21:04.377771Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 2638, "message_tokens": 2545, "message_utf8_bytes": 6643, "cache_tensor_bytes": null, "serialized_payload_bytes": 6643, "side_information_bytes": 0, "fetched_bytes": 3548, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25926.604749984108, "ttft_ms": 13.791457982733846, "decode_ms": 108.8209169683978, "receiver_prefill_ms": 314.0265420079231}, "total_wall_ms": 26366.75266700331, "ttft_ms": 13.791457982733846, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "19787467", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S3"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ea82bca55b3e0395c0cbaeda4a4f3918e461d53edcf7cf50f7f693eda7725b6d", "output_hash": "cce18353e296801e52ba93f1620a547b8c476782dc055877302afd6b9c41094c"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "13f65c1f883c64205b52a6ba", "attempt_id": "057f174e6b1e", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:21:04.835891Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 2600, "message_tokens": 2506, "message_utf8_bytes": 6595, "cache_tensor_bytes": null, "serialized_payload_bytes": 6595, "side_information_bytes": 0, "fetched_bytes": 3483, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25926.604749984108, "ttft_ms": 14.13370796944946, "decode_ms": 14.071166981011629, "receiver_prefill_ms": 371.43983296118677}, "total_wall_ms": 26329.926291015, "ttft_ms": 14.13370796944946, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S3"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "d9a5973b60ab3c3953ad8a714f347c08b38921c04aa0e60c43436a25bb970a37", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "4dece08794f4a0d446436a8b", "attempt_id": "671152b617c1", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:21:05.256633Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 2640, "message_tokens": 2545, "message_utf8_bytes": 6643, "cache_tensor_bytes": null, "serialized_payload_bytes": 6643, "side_information_bytes": 0, "fetched_bytes": 3548, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25926.604749984108, "ttft_ms": 13.777249958366156, "decode_ms": 13.664625003002584, "receiver_prefill_ms": 326.12950005568564}, "total_wall_ms": 26283.66204199847, "ttft_ms": 13.777249958366156, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S3"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ae26344b0810cddf24f68f62f78d670912b066c64a2038470cc2211ec60098a5", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "add57b88eb31fd675fd20c9d", "attempt_id": "504a28580a2c", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:21:06.649271Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 234749952}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 44, "message_tokens": null, "message_utf8_bytes": 30674, "cache_tensor_bytes": 234749952, "serialized_payload_bytes": 234749952, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 1017.1460410347208, "ttft_ms": 47.9392089182511, "decode_ms": 137.4198340345174, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1209.9817910930142, "ttft_ms": 47.9392089182511, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8,832,117", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S3"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "5433ad7dca6601c11c8a49e3239b880460c2fb281eb569df6a6134b5610e2859"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "3864797d1e5437366a33a084", "attempt_id": "69039880827f", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:21:06.860180Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 234749952}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 46, "message_tokens": null, "message_utf8_bytes": 30674, "cache_tensor_bytes": 234749952, "serialized_payload_bytes": 234749952, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 1017.1460410347208, "ttft_ms": 112.61112499050796, "decode_ms": 151.60129207652062, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1285.028374986723, "ttft_ms": 112.61112499050796, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "RY-99D-9177", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S3"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "abc5e1adc6533afa7ee1683a2195252eb090b08b52392c8150ccf6a387b14df6"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "2fed6b97c0d4cead6e29072b", "attempt_id": "277bb14922cf", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:21:07.145867Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 234749952}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 52, "message_tokens": null, "message_utf8_bytes": 30674, "cache_tensor_bytes": 234749952, "serialized_payload_bytes": 234749952, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 1017.1460410347208, "ttft_ms": 96.22266597580165, "decode_ms": 105.50725006032735, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1223.503583110869, "ttft_ms": 96.22266597580165, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7755450", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S3"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "a30221bc7c855d26c2bda21a0a21fda10c953a05d985cec4e1dfa5a6a372d5d3"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "8db54b6326bd2a44692d59b7", "attempt_id": "d9bcca6ce549", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:21:07.370283Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 234749952}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 45, "message_tokens": null, "message_utf8_bytes": 30674, "cache_tensor_bytes": 234749952, "serialized_payload_bytes": 234749952, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 1017.1460410347208, "ttft_ms": 77.41875003557652, "decode_ms": 135.72104193735868, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1234.8511239979416, "ttft_ms": 77.41875003557652, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7,430,110", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S3"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "6ec8f3686ca12e2212d5f0b03da40f6c90866d495314345bafd9be7b9638f772"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "0b75611016609253bdbee20a", "attempt_id": "57f500b1295e", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:21:07.605694Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 234749952}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 54, "message_tokens": null, "message_utf8_bytes": 30674, "cache_tensor_bytes": 234749952, "serialized_payload_bytes": 234749952, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 1017.1460410347208, "ttft_ms": 94.24754197243601, "decode_ms": 105.74304091278464, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1221.6917909681797, "ttft_ms": 94.24754197243601, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Lithmunpra-16", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S3"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "26fc169ceaba84d7f292defd76062949b1b87a9a87be89b5c857b8642cc75061"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "75de46ada3b09a58894ee44f", "attempt_id": "f51eb890eadb", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:21:07.827904Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 234749952}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 30674, "cache_tensor_bytes": 234749952, "serialized_payload_bytes": 234749952, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 1017.1460410347208, "ttft_ms": 85.24237503297627, "decode_ms": 124.29979210719466, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1231.2796240439638, "ttft_ms": 85.24237503297627, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "11528867", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S3"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "ca18c17344e52eec5efdd8dcbd48ea4efa05565b96708c06fbbc020aaefd830d"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "f2055a88b6d29fff164e7e3e", "attempt_id": "68bb869e6c50", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:21:08.060018Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 234749952}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 48, "message_tokens": null, "message_utf8_bytes": 30674, "cache_tensor_bytes": 234749952, "serialized_payload_bytes": 234749952, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 1017.1460410347208, "ttft_ms": 46.566916978918016, "decode_ms": 15.094375004991889, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1082.4113739654422, "ttft_ms": 46.566916978918016, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S3"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "6a300ba49d4d543710025a66", "attempt_id": "27abf8edc7dc", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:21:08.143232Z", "task_family": "synthetic", "document_id": "syn_03016_8000", "document_hash": "c087f3475bd3d0ecdd3f33c8a5862c81d3ce40f58b9774fe26f7e5627786e04f", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03016_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 234749952}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5802, "query_tokens": 49, "message_tokens": null, "message_utf8_bytes": 30674, "cache_tensor_bytes": 234749952, "serialized_payload_bytes": 234749952, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 1017.1460410347208, "ttft_ms": 87.73899998050183, "decode_ms": 15.255582984536886, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1124.6621659956872, "ttft_ms": 87.73899998050183, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Ixtek-9", "bogus_amount": "3,686,620", "source": "S3"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03016_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "6201d2e1de696820a123fd32", "attempt_id": "15b8c8843424", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:21:09.240541Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 2367, "message_tokens": 2280, "message_utf8_bytes": 8213, "cache_tensor_bytes": null, "serialized_payload_bytes": 8213, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.03783300053328276, "ttft_ms": 14.085083967074752, "decode_ms": 123.31195792648941, "receiver_prefill_ms": 271.97479200549424}, "total_wall_ms": 415.1403739815578, "ttft_ms": 14.085083967074752, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7,296,125", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S1"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "f5e6c9a75a40744953e5a3065ae2109222bc2193820d02f1d88371a038846af7", "output_hash": "8b531ccf7fd3b97bf2ef32627da2617b8353e5c2dd09a274017a5a5a868bdcdb"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "5ca9f4fead8b258599df6c58", "attempt_id": "82a706cf7fbc", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:21:09.669788Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 2367, "message_tokens": 2280, "message_utf8_bytes": 8213, "cache_tensor_bytes": null, "serialized_payload_bytes": 8213, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.03783300053328276, "ttft_ms": 13.415875029750168, "decode_ms": 135.6212500249967, "receiver_prefill_ms": 259.5347079914063}, "total_wall_ms": 412.23545803222805, "ttft_ms": 13.415875029750168, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "HM-38K-5533", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S1"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "f5e6c9a75a40744953e5a3065ae2109222bc2193820d02f1d88371a038846af7", "output_hash": "18acd9807857243bbded3bceae55398316a3958b81d70364441bdb6748eaf878"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "dd566a2b4f1ca15e9aae77a3", "attempt_id": "eb350fc3ec9d", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:21:10.096437Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 2376, "message_tokens": 2280, "message_utf8_bytes": 8213, "cache_tensor_bytes": null, "serialized_payload_bytes": 8213, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.03783300053328276, "ttft_ms": 13.736417051404715, "decode_ms": 95.44350008945912, "receiver_prefill_ms": 261.89208298455924}, "total_wall_ms": 374.949041986838, "ttft_ms": 13.736417051404715, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "276,340", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S1"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "f5e6c9a75a40744953e5a3065ae2109222bc2193820d02f1d88371a038846af7", "output_hash": "ef0dffffbd6c962fe44611aa2052fb7b78efaa1203b0a5de9937b8bd584d71c2"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "5458ec33fce9ab77396b8d56", "attempt_id": "48b5fca42a0d", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:21:10.486236Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 2369, "message_tokens": 2280, "message_utf8_bytes": 8213, "cache_tensor_bytes": null, "serialized_payload_bytes": 8213, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.03783300053328276, "ttft_ms": 13.442374998703599, "decode_ms": 94.9155839625746, "receiver_prefill_ms": 258.7415830930695}, "total_wall_ms": 370.7315409556031, "ttft_ms": 13.442374998703599, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "283,475", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S1"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "f5e6c9a75a40744953e5a3065ae2109222bc2193820d02f1d88371a038846af7", "output_hash": "98c08bf99e13b01a52116e9ac4adbfbe6ef32ccac770135a0983cd88f756fadf"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "bd01ee805dc40ffad72cb689", "attempt_id": "e9b5d0423f75", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:21:10.871670Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 2377, "message_tokens": 2280, "message_utf8_bytes": 8213, "cache_tensor_bytes": null, "serialized_payload_bytes": 8213, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.03783300053328276, "ttft_ms": 13.540958054363728, "decode_ms": 81.38320792932063, "receiver_prefill_ms": 261.28737500403076}, "total_wall_ms": 359.81937404721975, "ttft_ms": 13.540958054363728, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Melymi-15", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S1"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "f5e6c9a75a40744953e5a3065ae2109222bc2193820d02f1d88371a038846af7", "output_hash": "fa08d1474b01d6dfbf3dea049be251aa3c3021c17b4818258064b156509af58b"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "6e5312f4c22df3fb5724644d", "attempt_id": "ec7b42388bb8", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:21:11.245959Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 2369, "message_tokens": 2280, "message_utf8_bytes": 8213, "cache_tensor_bytes": null, "serialized_payload_bytes": 8213, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.03783300053328276, "ttft_ms": 13.577334000729024, "decode_ms": 108.24604192748666, "receiver_prefill_ms": 258.5030000191182}, "total_wall_ms": 383.9046669891104, "ttft_ms": 13.577334000729024, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "13034900", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S1"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "f5e6c9a75a40744953e5a3065ae2109222bc2193820d02f1d88371a038846af7", "output_hash": "7452f48d4acaf50737972dd0f28b0c8312ce696c5915754b46408a89c313061d"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "89503ff2601af66be8752774", "attempt_id": "7e6033766de5", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:21:11.644240Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 2370, "message_tokens": 2280, "message_utf8_bytes": 8213, "cache_tensor_bytes": null, "serialized_payload_bytes": 8213, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.03783300053328276, "ttft_ms": 13.583499938249588, "decode_ms": 13.734125066548586, "receiver_prefill_ms": 261.3877080148086}, "total_wall_ms": 292.3034990672022, "ttft_ms": 13.583499938249588, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S1"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "f5e6c9a75a40744953e5a3065ae2109222bc2193820d02f1d88371a038846af7", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "85593f38ad9e1f6c11f5253d", "attempt_id": "ccd264750255", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:21:11.950486Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 2371, "message_tokens": 2280, "message_utf8_bytes": 8213, "cache_tensor_bytes": null, "serialized_payload_bytes": 8213, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.03783300053328276, "ttft_ms": 13.47025006543845, "decode_ms": 13.587917084805667, "receiver_prefill_ms": 261.61687506828457}, "total_wall_ms": 292.221250012517, "ttft_ms": 13.47025006543845, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S1"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "f5e6c9a75a40744953e5a3065ae2109222bc2193820d02f1d88371a038846af7", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "caa43d7637acc826adb7949b", "attempt_id": "c0b0539de617", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:21:26.594284Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 1125, "message_tokens": 1037, "message_utf8_bytes": 2486, "cache_tensor_bytes": null, "serialized_payload_bytes": 2486, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14337.31179102324, "ttft_ms": 14.954000012949109, "decode_ms": 95.25350003968924, "receiver_prefill_ms": 121.89750000834465}, "total_wall_ms": 14571.948624099605, "ttft_ms": 14.954000012949109, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8174700", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S1"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "1c71f9fc635d84a54851baecd73364ea0f91b7583ac9040e8a1f4e5d0c6722f7", "output_hash": "4ee6eeeb1e8e5eda62b03599d84aad91f3744be1bef7d10a0ffbd2ffed71dc72"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "72076c61adda363446d06485", "attempt_id": "7817a4ed0e48", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:21:26.844867Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 1125, "message_tokens": 1037, "message_utf8_bytes": 2486, "cache_tensor_bytes": null, "serialized_payload_bytes": 2486, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14337.31179102324, "ttft_ms": 13.452125014737248, "decode_ms": 138.8915419811383, "receiver_prefill_ms": 121.06991605833173}, "total_wall_ms": 14612.70862503443, "ttft_ms": 13.452125014737248, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "HM-38K-5533", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S1"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "1c71f9fc635d84a54851baecd73364ea0f91b7583ac9040e8a1f4e5d0c6722f7", "output_hash": "18acd9807857243bbded3bceae55398316a3958b81d70364441bdb6748eaf878"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "4752b285f37dd5c1e2ff6e65", "attempt_id": "6780487a1c1a", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:21:27.134224Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 1134, "message_tokens": 1037, "message_utf8_bytes": 2486, "cache_tensor_bytes": null, "serialized_payload_bytes": 2486, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14337.31179102324, "ttft_ms": 13.467375072650611, "decode_ms": 94.21979205217212, "receiver_prefill_ms": 121.35720800142735}, "total_wall_ms": 14568.269707029685, "ttft_ms": 13.467375072650611, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "276,340", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S1"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "1c71f9fc635d84a54851baecd73364ea0f91b7583ac9040e8a1f4e5d0c6722f7", "output_hash": "ef0dffffbd6c962fe44611aa2052fb7b78efaa1203b0a5de9937b8bd584d71c2"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "e2323b19a71c0b0d8a4cea91", "attempt_id": "95bdde0304ef", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:21:27.380446Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 1127, "message_tokens": 1037, "message_utf8_bytes": 2486, "cache_tensor_bytes": null, "serialized_payload_bytes": 2486, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14337.31179102324, "ttft_ms": 14.071541954763234, "decode_ms": 81.14366710651666, "receiver_prefill_ms": 121.57945800572634}, "total_wall_ms": 14556.021749973297, "ttft_ms": 14.071541954763234, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "304815", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S1"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "1c71f9fc635d84a54851baecd73364ea0f91b7583ac9040e8a1f4e5d0c6722f7", "output_hash": "a480475d6855141c920f42d1bf3337ca7fd88bcf4087b0ed2e253ed68cca6926"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "22519f631bcae8ca7d037647", "attempt_id": "e87e36d605f9", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:21:27.613743Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 1135, "message_tokens": 1037, "message_utf8_bytes": 2486, "cache_tensor_bytes": null, "serialized_payload_bytes": 2486, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14337.31179102324, "ttft_ms": 13.863958069123328, "decode_ms": 81.28649997524917, "receiver_prefill_ms": 121.46679195575416}, "total_wall_ms": 14555.80875009764, "ttft_ms": 13.863958069123328, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Melymi-15", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S1"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "1c71f9fc635d84a54851baecd73364ea0f91b7583ac9040e8a1f4e5d0c6722f7", "output_hash": "fa08d1474b01d6dfbf3dea049be251aa3c3021c17b4818258064b156509af58b"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "6d7d96356ac0716ea38f0d44", "attempt_id": "11c6dd3933fc", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:21:27.846752Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 1127, "message_tokens": 1037, "message_utf8_bytes": 2486, "cache_tensor_bytes": null, "serialized_payload_bytes": 2486, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14337.31179102324, "ttft_ms": 13.999832910485566, "decode_ms": 97.10545896086842, "receiver_prefill_ms": 122.70224990788847}, "total_wall_ms": 14573.0577489594, "ttft_ms": 13.999832910485566, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1544061", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S1"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "1c71f9fc635d84a54851baecd73364ea0f91b7583ac9040e8a1f4e5d0c6722f7", "output_hash": "e191fd75967d97432a8139db7a118a20e82948405d2e3fb5dce4e7ce74a2ca7f"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c78a1acf2fc395e9c90df712", "attempt_id": "a9c5b203d7bc", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:21:28.096987Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 1128, "message_tokens": 1037, "message_utf8_bytes": 2486, "cache_tensor_bytes": null, "serialized_payload_bytes": 2486, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14337.31179102324, "ttft_ms": 13.544375076889992, "decode_ms": 13.779375003650784, "receiver_prefill_ms": 128.73866606969386}, "total_wall_ms": 14495.35208195448, "ttft_ms": 13.544375076889992, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S1"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "1c71f9fc635d84a54851baecd73364ea0f91b7583ac9040e8a1f4e5d0c6722f7", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "6f5a2b55903c3f2e21dc8019", "attempt_id": "a7e3df8ff982", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:21:28.270328Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 1129, "message_tokens": 1037, "message_utf8_bytes": 2486, "cache_tensor_bytes": null, "serialized_payload_bytes": 2486, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14337.31179102324, "ttft_ms": 14.085708069615066, "decode_ms": 13.712375075556338, "receiver_prefill_ms": 133.7098750518635}, "total_wall_ms": 14500.75220700819, "ttft_ms": 14.085708069615066, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S1"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "1c71f9fc635d84a54851baecd73364ea0f91b7583ac9040e8a1f4e5d0c6722f7", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "4ea935e9fc0549bb989d8a95", "attempt_id": "f853e5a74031", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:21:41.681787Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 14}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 665, "message_tokens": 573, "message_utf8_bytes": 1496, "cache_tensor_bytes": null, "serialized_payload_bytes": 1496, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13233.806040952913, "ttft_ms": 13.468958088196814, "decode_ms": 13.337875017896295, "receiver_prefill_ms": 75.63533401116729}, "total_wall_ms": 13340.643457951955, "ttft_ms": 13.468958088196814, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S1"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "974f1a0db52d7ba890c828f941d92c7320f20fd4cee4992caaacbaafdaeec0cd", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "0070618d9809764d004998d3", "attempt_id": "83cb92732b29", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:21:41.803000Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 14}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 665, "message_tokens": 573, "message_utf8_bytes": 1496, "cache_tensor_bytes": null, "serialized_payload_bytes": 1496, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13233.806040952913, "ttft_ms": 13.24154098983854, "decode_ms": 13.578833895735443, "receiver_prefill_ms": 74.64616594370455}, "total_wall_ms": 13336.572416010313, "ttft_ms": 13.24154098983854, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S1"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "974f1a0db52d7ba890c828f941d92c7320f20fd4cee4992caaacbaafdaeec0cd", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b5135367e8bdf968e31b97b4", "attempt_id": "f99d76a44d65", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:21:41.919928Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 14}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 674, "message_tokens": 573, "message_utf8_bytes": 1496, "cache_tensor_bytes": null, "serialized_payload_bytes": 1496, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13233.806040952913, "ttft_ms": 13.24920798651874, "decode_ms": 80.20074991509318, "receiver_prefill_ms": 76.69912499841303}, "total_wall_ms": 13405.308498884551, "ttft_ms": 13.24920798651874, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "276340", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S1"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "974f1a0db52d7ba890c828f941d92c7320f20fd4cee4992caaacbaafdaeec0cd", "output_hash": "bc7c3667cbe0ad74ec0a1b092d3bad7e6a02978fa67746e3adbf3174ed34b3a2"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "d51102e6e59ec475df6c8f60", "attempt_id": "742b2e27b8f5", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:21:42.105742Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 14}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 667, "message_tokens": 573, "message_utf8_bytes": 1496, "cache_tensor_bytes": null, "serialized_payload_bytes": 1496, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13233.806040952913, "ttft_ms": 13.385790982283652, "decode_ms": 80.07799996994436, "receiver_prefill_ms": 75.15658298507333}, "total_wall_ms": 13403.7407488795, "ttft_ms": 13.385790982283652, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "276340", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S1"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "974f1a0db52d7ba890c828f941d92c7320f20fd4cee4992caaacbaafdaeec0cd", "output_hash": "bc7c3667cbe0ad74ec0a1b092d3bad7e6a02978fa67746e3adbf3174ed34b3a2"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "1f745e0e4e10d4eb6bf00292", "attempt_id": "c8f21801cdaa", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:21:42.290502Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 14}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 675, "message_tokens": 573, "message_utf8_bytes": 1496, "cache_tensor_bytes": null, "serialized_payload_bytes": 1496, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13233.806040952913, "ttft_ms": 13.216332998126745, "decode_ms": 79.62670805864036, "receiver_prefill_ms": 76.06766698881984}, "total_wall_ms": 13404.037540894933, "ttft_ms": 13.216332998126745, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Ruknov-12", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S1"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "974f1a0db52d7ba890c828f941d92c7320f20fd4cee4992caaacbaafdaeec0cd", "output_hash": "f7aae1e5b23999fd5f007dabe7ac59df11265b20b68c84615515729028871151"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a6d576888796a211ac4fa658", "attempt_id": "ec61b6b83acb", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:21:42.475348Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 14}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 667, "message_tokens": 573, "message_utf8_bytes": 1496, "cache_tensor_bytes": null, "serialized_payload_bytes": 1496, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13233.806040952913, "ttft_ms": 13.669500011019409, "decode_ms": 80.28250001370907, "receiver_prefill_ms": 74.1992499679327}, "total_wall_ms": 13403.31983298529, "ttft_ms": 13.669500011019409, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "272184", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S1"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "974f1a0db52d7ba890c828f941d92c7320f20fd4cee4992caaacbaafdaeec0cd", "output_hash": "3aadaddf09c30b8bfe5dc7e2aa6984fcf0df4d216145cf925760dd5a946d6e05"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "5fbaf594137fc8bb0432cb14", "attempt_id": "049c43fd050d", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:21:42.659077Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 14}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 668, "message_tokens": 573, "message_utf8_bytes": 1496, "cache_tensor_bytes": null, "serialized_payload_bytes": 1496, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13233.806040952913, "ttft_ms": 13.287291978485882, "decode_ms": 13.374665984883904, "receiver_prefill_ms": 74.9332089908421}, "total_wall_ms": 13336.708706920035, "ttft_ms": 13.287291978485882, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S1"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "974f1a0db52d7ba890c828f941d92c7320f20fd4cee4992caaacbaafdaeec0cd", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "6ad3446bb9cc746488621182", "attempt_id": "964922ac6ae6", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:21:42.776681Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 14}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 669, "message_tokens": 573, "message_utf8_bytes": 1496, "cache_tensor_bytes": null, "serialized_payload_bytes": 1496, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13233.806040952913, "ttft_ms": 13.521374901756644, "decode_ms": 13.765749987214804, "receiver_prefill_ms": 74.53954091761261}, "total_wall_ms": 13336.957498919219, "ttft_ms": 13.521374901756644, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S1"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "974f1a0db52d7ba890c828f941d92c7320f20fd4cee4992caaacbaafdaeec0cd", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "843780852ddd589ecb9ba4fb", "attempt_id": "1d68d702896e", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:21:56.144146Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 775, "message_tokens": 684, "message_utf8_bytes": 1708, "cache_tensor_bytes": null, "serialized_payload_bytes": 1708, "side_information_bytes": 0, "fetched_bytes": 906, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13250.139041105285, "ttft_ms": 13.414249988272786, "decode_ms": 13.273083954118192, "receiver_prefill_ms": 92.23841701168567}, "total_wall_ms": 13370.58899900876, "ttft_ms": 13.414249988272786, "input_usage": {"lookups": 3}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S1"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "140567b9792f9def76c4658c8f811c9e455952cd833fc2d6b1f6a8e492f73051", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a77d3f659dbbc37d09157f53", "attempt_id": "aa3c76f97377", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:21:56.278882Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 775, "message_tokens": 684, "message_utf8_bytes": 1708, "cache_tensor_bytes": null, "serialized_payload_bytes": 1708, "side_information_bytes": 0, "fetched_bytes": 906, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13250.139041105285, "ttft_ms": 13.490915996953845, "decode_ms": 134.161000023596, "receiver_prefill_ms": 92.03012497164309}, "total_wall_ms": 13491.48850014899, "ttft_ms": 13.490915996953845, "input_usage": {"lookups": 3}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "HM-38K-5533", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S1"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "140567b9792f9def76c4658c8f811c9e455952cd833fc2d6b1f6a8e492f73051", "output_hash": "18acd9807857243bbded3bceae55398316a3958b81d70364441bdb6748eaf878"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "bb89d56674a3e15cc2345ce9", "attempt_id": "cf4722c5e325", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:21:56.535066Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 756, "message_tokens": 656, "message_utf8_bytes": 1693, "cache_tensor_bytes": null, "serialized_payload_bytes": 1693, "side_information_bytes": 0, "fetched_bytes": 888, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13250.139041105285, "ttft_ms": 13.482250040397048, "decode_ms": 80.22645907476544, "receiver_prefill_ms": 81.28608309198171}, "total_wall_ms": 13426.718791131862, "ttft_ms": 13.482250040397048, "input_usage": {"lookups": 3}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "283475", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S1"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "bba483cf0314a1950b107de635beee8cd8087e6fac20244258e47b2aaa9dbe64", "output_hash": "1b791433a3d20cf3e6d069df1e033eee1855e3d410e707882c3854696e18a3a0"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "95473bc1fc425afc52865cfb", "attempt_id": "3f362237dcc9", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:21:56.725662Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 749, "message_tokens": 656, "message_utf8_bytes": 1693, "cache_tensor_bytes": null, "serialized_payload_bytes": 1693, "side_information_bytes": 0, "fetched_bytes": 888, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13250.139041105285, "ttft_ms": 13.27633298933506, "decode_ms": 80.42258396744728, "receiver_prefill_ms": 82.03112508635968}, "total_wall_ms": 13427.326749195345, "ttft_ms": 13.27633298933506, "input_usage": {"lookups": 3}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "283475", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S1"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "bba483cf0314a1950b107de635beee8cd8087e6fac20244258e47b2aaa9dbe64", "output_hash": "1b791433a3d20cf3e6d069df1e033eee1855e3d410e707882c3854696e18a3a0"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "58952951a6a1ce258ad523a0", "attempt_id": "ad42dc31f938", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:21:56.917160Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 787, "message_tokens": 686, "message_utf8_bytes": 1697, "cache_tensor_bytes": null, "serialized_payload_bytes": 1697, "side_information_bytes": 0, "fetched_bytes": 897, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13250.139041105285, "ttft_ms": 13.441874994896352, "decode_ms": 80.29650000389665, "receiver_prefill_ms": 92.08724997006357}, "total_wall_ms": 13438.626041053794, "ttft_ms": 13.441874994896352, "input_usage": {"lookups": 3}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Naktisvar-5", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S1"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a8c3800e34062527d96990b043c2a741138192717a152fa7868d2d5f278a9978", "output_hash": "6c1697b92554c5222d4614ccbd3c5dd33e97f20b46dd14b6eaff4e66238d1cf1"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "6f80e8c82e9d14cd9e0ff5f1", "attempt_id": "22f26dae3ead", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:21:57.120777Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 777, "message_tokens": 684, "message_utf8_bytes": 1708, "cache_tensor_bytes": null, "serialized_payload_bytes": 1708, "side_information_bytes": 0, "fetched_bytes": 906, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13250.139041105285, "ttft_ms": 13.304749969393015, "decode_ms": 107.50445898156613, "receiver_prefill_ms": 91.94262500386685}, "total_wall_ms": 13464.358333148994, "ttft_ms": 13.304749969393015, "input_usage": {"lookups": 3}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "13389786", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S1"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "140567b9792f9def76c4658c8f811c9e455952cd833fc2d6b1f6a8e492f73051", "output_hash": "c5e806e1430a6ff86bd0d91093bbe7324a41241ff6c526b61fd13e0997b35011"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "8fc46ec9f34bfd8c253c197d", "attempt_id": "586b89333ffb", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:21:57.349146Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 778, "message_tokens": 684, "message_utf8_bytes": 1708, "cache_tensor_bytes": null, "serialized_payload_bytes": 1708, "side_information_bytes": 0, "fetched_bytes": 906, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13250.139041105285, "ttft_ms": 13.450709055177867, "decode_ms": 13.40329204685986, "receiver_prefill_ms": 92.23525004927069}, "total_wall_ms": 13370.721541112289, "ttft_ms": 13.450709055177867, "input_usage": {"lookups": 3}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S1"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "140567b9792f9def76c4658c8f811c9e455952cd833fc2d6b1f6a8e492f73051", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "0ba9cb325a2923aa658fc5c7", "attempt_id": "2c59e94b6e5f", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:21:57.484268Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 779, "message_tokens": 684, "message_utf8_bytes": 1708, "cache_tensor_bytes": null, "serialized_payload_bytes": 1708, "side_information_bytes": 0, "fetched_bytes": 906, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 13250.139041105285, "ttft_ms": 13.360208016820252, "decode_ms": 13.791250064969063, "receiver_prefill_ms": 91.92058397457004}, "total_wall_ms": 13370.710124145262, "ttft_ms": 13.360208016820252, "input_usage": {"lookups": 3}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S1"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "140567b9792f9def76c4658c8f811c9e455952cd833fc2d6b1f6a8e492f73051", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "f45c286b42961fa855ebed55", "attempt_id": "32f940e3d5cf", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:21:57.878162Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 86925312}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 45, "message_tokens": null, "message_utf8_bytes": 8187, "cache_tensor_bytes": 86925312, "serialized_payload_bytes": 86925312, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 259.060375043191, "ttft_ms": 28.488541953265667, "decode_ms": 97.37912495620549, "receiver_prefill_ms": 0.0}, "total_wall_ms": 389.1927500953898, "ttft_ms": 28.488541953265667, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7296125", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S1"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "9113f9982374a4745ee6d0ebc61cd26058453b2fe703d9445ac1818cb0f5005a"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "84fbfe81137b2601d692ff8e", "attempt_id": "d80d7a380f61", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:21:58.025650Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 86925312}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 45, "message_tokens": null, "message_utf8_bytes": 8187, "cache_tensor_bytes": 86925312, "serialized_payload_bytes": 86925312, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 259.060375043191, "ttft_ms": 28.296333039179444, "decode_ms": 143.34524993319064, "receiver_prefill_ms": 0.0}, "total_wall_ms": 434.6491670003161, "ttft_ms": 28.296333039179444, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "HY-60D-6477", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S1"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "594ae9f26be6c27dcc32dccfb480f5a02fccb94750108a965d3c8e60cf5c19d2"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "6140a917592b50e19b83c1e0", "attempt_id": "27df5f0ab46b", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:21:58.215911Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 86925312}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 54, "message_tokens": null, "message_utf8_bytes": 8187, "cache_tensor_bytes": 86925312, "serialized_payload_bytes": 86925312, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 259.060375043191, "ttft_ms": 27.946999995037913, "decode_ms": 84.02304106857628, "receiver_prefill_ms": 0.0}, "total_wall_ms": 375.0055411364883, "ttft_ms": 27.946999995037913, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "304815", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S1"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "a480475d6855141c920f42d1bf3337ca7fd88bcf4087b0ed2e253ed68cca6926"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "0cc2644463deced379bdd587", "attempt_id": "edf9c9c81ca1", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:21:58.346211Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 86925312}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 8187, "cache_tensor_bytes": 86925312, "serialized_payload_bytes": 86925312, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 259.060375043191, "ttft_ms": 28.742625028826296, "decode_ms": 85.71729203686118, "receiver_prefill_ms": 0.0}, "total_wall_ms": 377.46245809830725, "ttft_ms": 28.742625028826296, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "304815", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S1"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "a480475d6855141c920f42d1bf3337ca7fd88bcf4087b0ed2e253ed68cca6926"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "6c3aab0e5173f1fb10283b52", "attempt_id": "fb63293e290b", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:21:58.479095Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 86925312}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 55, "message_tokens": null, "message_utf8_bytes": 8187, "cache_tensor_bytes": 86925312, "serialized_payload_bytes": 86925312, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 259.060375043191, "ttft_ms": 28.948542079888284, "decode_ms": 83.52908294182271, "receiver_prefill_ms": 0.0}, "total_wall_ms": 374.43187506869435, "ttft_ms": 28.948542079888284, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Melymi-15", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S1"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "fa08d1474b01d6dfbf3dea049be251aa3c3021c17b4818258064b156509af58b"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "cd44f3995b3da3882fdf7cf1", "attempt_id": "4992848c6f94", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:21:58.608518Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 86925312}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 8187, "cache_tensor_bytes": 86925312, "serialized_payload_bytes": 86925312, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 259.060375043191, "ttft_ms": 28.717583976686, "decode_ms": 98.36908301804215, "receiver_prefill_ms": 0.0}, "total_wall_ms": 390.10645798407495, "ttft_ms": 28.717583976686, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5444459", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S1"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "177e1bba8b16b05766bb7b56dc68cec8185efbd6bd8825f766e1ade4f079a0c3"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "75423e22d3178a776b266aaf", "attempt_id": "23ea0c68a889", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:21:58.753867Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 86925312}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 48, "message_tokens": null, "message_utf8_bytes": 8187, "cache_tensor_bytes": 86925312, "serialized_payload_bytes": 86925312, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 259.060375043191, "ttft_ms": 28.913125046528876, "decode_ms": 13.85595800820738, "receiver_prefill_ms": 0.0}, "total_wall_ms": 305.74325006455183, "ttft_ms": 28.913125046528876, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S1"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "6a6cdb3b3123c2aa27582bb7", "attempt_id": "91481d377393", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:21:58.814287Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 86925312}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 49, "message_tokens": null, "message_utf8_bytes": 8187, "cache_tensor_bytes": 86925312, "serialized_payload_bytes": 86925312, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 259.060375043191, "ttft_ms": 28.589167050085962, "decode_ms": 13.801874942146242, "receiver_prefill_ms": 0.0}, "total_wall_ms": 305.3891670424491, "ttft_ms": 28.589167050085962, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S1"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b0ab82d4c9dc6636fb272608", "attempt_id": "b4bf961d610f", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:21:58.874489Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 2481, "message_tokens": 2394, "message_utf8_bytes": 8717, "cache_tensor_bytes": null, "serialized_payload_bytes": 8717, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.02683396451175213, "ttft_ms": 13.549875002354383, "decode_ms": 121.97341700084507, "receiver_prefill_ms": 270.4307499807328}, "total_wall_ms": 409.6873339731246, "ttft_ms": 13.549875002354383, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7,296,125", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S3"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "215244a1a6a853a67bc6380df092852f595d7b117cfe10fa2f976102dc032f9e", "output_hash": "8b531ccf7fd3b97bf2ef32627da2617b8353e5c2dd09a274017a5a5a868bdcdb"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "73d2c1c1ad0e3b280a32e121", "attempt_id": "282fc5e3e4ef", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:21:59.298152Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 2481, "message_tokens": 2394, "message_utf8_bytes": 8717, "cache_tensor_bytes": null, "serialized_payload_bytes": 8717, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.02683396451175213, "ttft_ms": 13.479207991622388, "decode_ms": 135.76599990483373, "receiver_prefill_ms": 268.3525829343125}, "total_wall_ms": 421.32195888552815, "ttft_ms": 13.479207991622388, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "HM-38K-5533", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S3"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "215244a1a6a853a67bc6380df092852f595d7b117cfe10fa2f976102dc032f9e", "output_hash": "18acd9807857243bbded3bceae55398316a3958b81d70364441bdb6748eaf878"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "d819d7f235e69be2af4602a6", "attempt_id": "16075ffce134", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:21:59.733917Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 2490, "message_tokens": 2394, "message_utf8_bytes": 8717, "cache_tensor_bytes": null, "serialized_payload_bytes": 8717, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.02683396451175213, "ttft_ms": 13.711792067624629, "decode_ms": 96.0936249466613, "receiver_prefill_ms": 274.6290420182049}, "total_wall_ms": 388.2901679025963, "ttft_ms": 13.711792067624629, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "276,340", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S3"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "215244a1a6a853a67bc6380df092852f595d7b117cfe10fa2f976102dc032f9e", "output_hash": "ef0dffffbd6c962fe44611aa2052fb7b78efaa1203b0a5de9937b8bd584d71c2"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "ac4be557427c886e3a56fe0e", "attempt_id": "25409d070af6", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:22:00.136151Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 2483, "message_tokens": 2394, "message_utf8_bytes": 8717, "cache_tensor_bytes": null, "serialized_payload_bytes": 8717, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.02683396451175213, "ttft_ms": 13.912333059124649, "decode_ms": 95.67895799409598, "receiver_prefill_ms": 289.4833330065012}, "total_wall_ms": 402.86129293963313, "ttft_ms": 13.912333059124649, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "276,340", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S3"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "215244a1a6a853a67bc6380df092852f595d7b117cfe10fa2f976102dc032f9e", "output_hash": "ef0dffffbd6c962fe44611aa2052fb7b78efaa1203b0a5de9937b8bd584d71c2"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "7faab244e63abaab317b919c", "attempt_id": "ebefc32afcfa", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:22:00.553796Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 2491, "message_tokens": 2394, "message_utf8_bytes": 8717, "cache_tensor_bytes": null, "serialized_payload_bytes": 8717, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.02683396451175213, "ttft_ms": 13.746500015258789, "decode_ms": 81.78250002674758, "receiver_prefill_ms": 287.6750829163939}, "total_wall_ms": 387.01458403375, "ttft_ms": 13.746500015258789, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Melymi-15", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S3"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "215244a1a6a853a67bc6380df092852f595d7b117cfe10fa2f976102dc032f9e", "output_hash": "fa08d1474b01d6dfbf3dea049be251aa3c3021c17b4818258064b156509af58b"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "dae97ca99b131dcb3d6f3065", "attempt_id": "daa737fe555b", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:22:00.954892Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 2483, "message_tokens": 2394, "message_utf8_bytes": 8717, "cache_tensor_bytes": null, "serialized_payload_bytes": 8717, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.02683396451175213, "ttft_ms": 13.65970796905458, "decode_ms": 109.48195797391236, "receiver_prefill_ms": 289.6004170179367}, "total_wall_ms": 416.50345898233354, "ttft_ms": 13.65970796905458, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "10044855", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S3"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "215244a1a6a853a67bc6380df092852f595d7b117cfe10fa2f976102dc032f9e", "output_hash": "225e45b8bdf9a3da51732a495842c32b2d5fcfd03c409b396a00e40e11d5dbaf"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "0e934f40dbd8ad4345b4aae5", "attempt_id": "76f4cb35519b", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:22:01.385935Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 2484, "message_tokens": 2394, "message_utf8_bytes": 8717, "cache_tensor_bytes": null, "serialized_payload_bytes": 8717, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.02683396451175213, "ttft_ms": 13.704916927963495, "decode_ms": 13.935625087469816, "receiver_prefill_ms": 284.20741693116724}, "total_wall_ms": 315.7291259849444, "ttft_ms": 13.704916927963495, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S3"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "215244a1a6a853a67bc6380df092852f595d7b117cfe10fa2f976102dc032f9e", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "0e829a8a31874b97ba204b9f", "attempt_id": "7a7173c401e0", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:22:01.716020Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 2485, "message_tokens": 2394, "message_utf8_bytes": 8717, "cache_tensor_bytes": null, "serialized_payload_bytes": 8717, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.02683396451175213, "ttft_ms": 13.70129210408777, "decode_ms": 13.730917009525001, "receiver_prefill_ms": 292.1572079649195}, "total_wall_ms": 323.3824169728905, "ttft_ms": 13.70129210408777, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S3"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "215244a1a6a853a67bc6380df092852f595d7b117cfe10fa2f976102dc032f9e", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b5245a1d0131e846c7daa14e", "attempt_id": "9ce311a0891f", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:22:26.839908Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1807}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 1920, "message_tokens": 1832, "message_utf8_bytes": 5969, "cache_tensor_bytes": null, "serialized_payload_bytes": 5969, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24786.072375020012, "ttft_ms": 13.42345797456801, "decode_ms": 121.77708302624524, "receiver_prefill_ms": 202.48737500514835}, "total_wall_ms": 25127.475833054632, "ttft_ms": 13.42345797456801, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7,296,125", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S3"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "b0e74d8a75de1d9115c706fe9c75a294966412c68af1c7ca20ac1f85611989cd", "output_hash": "8b531ccf7fd3b97bf2ef32627da2617b8353e5c2dd09a274017a5a5a868bdcdb"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "0ec39056297dfc6a6d35aef1", "attempt_id": "e1c7f226ee14", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:22:27.196202Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1807}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 1920, "message_tokens": 1832, "message_utf8_bytes": 5969, "cache_tensor_bytes": null, "serialized_payload_bytes": 5969, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24786.072375020012, "ttft_ms": 13.779249973595142, "decode_ms": 136.1835829447955, "receiver_prefill_ms": 201.53312501497567}, "total_wall_ms": 25140.47412504442, "ttft_ms": 13.779249973595142, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "HM-38K-5533", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S3"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "b0e74d8a75de1d9115c706fe9c75a294966412c68af1c7ca20ac1f85611989cd", "output_hash": "18acd9807857243bbded3bceae55398316a3958b81d70364441bdb6748eaf878"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "76f224fdbb0cd79ad0014d7e", "attempt_id": "758402e89f72", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:22:27.564505Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1807}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 1929, "message_tokens": 1832, "message_utf8_bytes": 5969, "cache_tensor_bytes": null, "serialized_payload_bytes": 5969, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24786.072375020012, "ttft_ms": 13.819791958667338, "decode_ms": 95.34674999304116, "receiver_prefill_ms": 204.06970800831914}, "total_wall_ms": 25102.265000110492, "ttft_ms": 13.819791958667338, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "276,340", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S3"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "b0e74d8a75de1d9115c706fe9c75a294966412c68af1c7ca20ac1f85611989cd", "output_hash": "ef0dffffbd6c962fe44611aa2052fb7b78efaa1203b0a5de9937b8bd584d71c2"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "5fff0d78c0e661316ddb84f9", "attempt_id": "3d23bb2efbfb", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:22:27.897852Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1807}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 1922, "message_tokens": 1832, "message_utf8_bytes": 5969, "cache_tensor_bytes": null, "serialized_payload_bytes": 5969, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24786.072375020012, "ttft_ms": 13.531583012081683, "decode_ms": 94.98158399946988, "receiver_prefill_ms": 204.4396250275895}, "total_wall_ms": 25102.006875094958, "ttft_ms": 13.531583012081683, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "272,184", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S3"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "b0e74d8a75de1d9115c706fe9c75a294966412c68af1c7ca20ac1f85611989cd", "output_hash": "7ab8ea46f050f78e243a876f64862b441080883dc7365da1e4abc893edc12fb0"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "55336486b9989deb9017c6af", "attempt_id": "6366fa182157", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:22:28.230866Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1807}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 1930, "message_tokens": 1832, "message_utf8_bytes": 5969, "cache_tensor_bytes": null, "serialized_payload_bytes": 5969, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24786.072375020012, "ttft_ms": 13.63729196600616, "decode_ms": 81.68345806188881, "receiver_prefill_ms": 206.04833296965808}, "total_wall_ms": 25090.453833108768, "ttft_ms": 13.63729196600616, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Melymi-15", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S3"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "b0e74d8a75de1d9115c706fe9c75a294966412c68af1c7ca20ac1f85611989cd", "output_hash": "fa08d1474b01d6dfbf3dea049be251aa3c3021c17b4818258064b156509af58b"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "7c9c6869499d1e7ce938ce93", "attempt_id": "1489805b3276", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:22:28.549557Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1807}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 1922, "message_tokens": 1832, "message_utf8_bytes": 5969, "cache_tensor_bytes": null, "serialized_payload_bytes": 5969, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24786.072375020012, "ttft_ms": 13.83562502451241, "decode_ms": 94.88958294969052, "receiver_prefill_ms": 218.64362503401935}, "total_wall_ms": 25116.38129199855, "ttft_ms": 13.83562502451241, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1014440", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S3"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "b0e74d8a75de1d9115c706fe9c75a294966412c68af1c7ca20ac1f85611989cd", "output_hash": "9ee07f2945b343117bbec8183e4896adba5c8b89d8986ddde98f05ba5c5acfe1"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a499824a5bb592ff7eadf62e", "attempt_id": "3d0ba9a7f164", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:22:28.894630Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1807}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 1923, "message_tokens": 1832, "message_utf8_bytes": 5969, "cache_tensor_bytes": null, "serialized_payload_bytes": 5969, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24786.072375020012, "ttft_ms": 13.800916960462928, "decode_ms": 13.768708915449679, "receiver_prefill_ms": 217.8993329871446}, "total_wall_ms": 25034.52012501657, "ttft_ms": 13.800916960462928, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S3"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "b0e74d8a75de1d9115c706fe9c75a294966412c68af1c7ca20ac1f85611989cd", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "319f9613c54660b65700ed7a", "attempt_id": "ffaddf70be09", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:22:29.157332Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1807}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 1924, "message_tokens": 1832, "message_utf8_bytes": 5969, "cache_tensor_bytes": null, "serialized_payload_bytes": 5969, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24786.072375020012, "ttft_ms": 13.932958012446761, "decode_ms": 13.872291892766953, "receiver_prefill_ms": 227.83470898866653}, "total_wall_ms": 25044.62945810519, "ttft_ms": 13.932958012446761, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S3"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "b0e74d8a75de1d9115c706fe9c75a294966412c68af1c7ca20ac1f85611989cd", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "3484d8042d4dd35825547b49", "attempt_id": "6679ef524f4c", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:22:53.505830Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 25}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 1095, "message_tokens": 1003, "message_utf8_bytes": 2591, "cache_tensor_bytes": null, "serialized_payload_bytes": 2591, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24075.56304091122, "ttft_ms": 13.388999970629811, "decode_ms": 93.91679195687175, "receiver_prefill_ms": 118.44220897182822}, "total_wall_ms": 24303.05983289145, "ttft_ms": 13.388999970629811, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7296125", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S3"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "6de94ce0a0382b5fcf2789f34d988c1249a9cec61fa53445fcb9c0d093778e0b", "output_hash": "9113f9982374a4745ee6d0ebc61cd26058453b2fe703d9445ac1818cb0f5005a"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "ce32cb9133ea2074a77f9b9f", "attempt_id": "8e2a59a341ce", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:22:53.748326Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 25}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 1095, "message_tokens": 1003, "message_utf8_bytes": 2591, "cache_tensor_bytes": null, "serialized_payload_bytes": 2591, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24075.56304091122, "ttft_ms": 13.303124927915633, "decode_ms": 13.230542070232332, "receiver_prefill_ms": 117.64191696420312}, "total_wall_ms": 24221.49754082784, "ttft_ms": 13.303124927915633, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S3"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "6de94ce0a0382b5fcf2789f34d988c1249a9cec61fa53445fcb9c0d093778e0b", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "5824307c7e052b21b09dc54c", "attempt_id": "dbde75df96bc", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:22:53.911232Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 25}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 1104, "message_tokens": 1003, "message_utf8_bytes": 2591, "cache_tensor_bytes": null, "serialized_payload_bytes": 2591, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24075.56304091122, "ttft_ms": 13.526999973692, "decode_ms": 80.69170895032585, "receiver_prefill_ms": 146.04716701433063}, "total_wall_ms": 24317.551331943832, "ttft_ms": 13.526999973692, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "276340", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S3"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "6de94ce0a0382b5fcf2789f34d988c1249a9cec61fa53445fcb9c0d093778e0b", "output_hash": "bc7c3667cbe0ad74ec0a1b092d3bad7e6a02978fa67746e3adbf3174ed34b3a2"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "6095feebf20dbee48bd327c8", "attempt_id": "4e840d845370", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:22:54.167652Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 25}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 1097, "message_tokens": 1003, "message_utf8_bytes": 2591, "cache_tensor_bytes": null, "serialized_payload_bytes": 2591, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24075.56304091122, "ttft_ms": 14.360249973833561, "decode_ms": 85.86074993945658, "receiver_prefill_ms": 897.0807500882074}, "total_wall_ms": 25074.660249985754, "ttft_ms": 14.360249973833561, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "276340", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S3"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "6de94ce0a0382b5fcf2789f34d988c1249a9cec61fa53445fcb9c0d093778e0b", "output_hash": "bc7c3667cbe0ad74ec0a1b092d3bad7e6a02978fa67746e3adbf3174ed34b3a2"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b1a375df65d87d7688919b77", "attempt_id": "4ffc317fe720", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:22:55.181480Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 25}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 1105, "message_tokens": 1003, "message_utf8_bytes": 2591, "cache_tensor_bytes": null, "serialized_payload_bytes": 2591, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24075.56304091122, "ttft_ms": 13.727666926570237, "decode_ms": 86.07316599227488, "receiver_prefill_ms": 677.9985839966685}, "total_wall_ms": 24855.52124888636, "ttft_ms": 13.727666926570237, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Melymi-15", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S3"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "6de94ce0a0382b5fcf2789f34d988c1249a9cec61fa53445fcb9c0d093778e0b", "output_hash": "fa08d1474b01d6dfbf3dea049be251aa3c3021c17b4818258064b156509af58b"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "87ad61f60c6f0e9eb41c78b2", "attempt_id": "8e1a1504c201", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:22:55.981736Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 25}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 1097, "message_tokens": 1003, "message_utf8_bytes": 2591, "cache_tensor_bytes": null, "serialized_payload_bytes": 2591, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24075.56304091122, "ttft_ms": 13.843583990819752, "decode_ms": 95.62641708180308, "receiver_prefill_ms": 118.31199994776398}, "total_wall_ms": 24305.33970799297, "ttft_ms": 13.843583990819752, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "2284681", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S3"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "6de94ce0a0382b5fcf2789f34d988c1249a9cec61fa53445fcb9c0d093778e0b", "output_hash": "827fd444ea8bf66041d71f150551d48316277f6e1ae79da829c800b60168ba6e"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "5f99be46c5faf5d75f563a60", "attempt_id": "f92a4ab4b14d", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:22:56.233290Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 25}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 1098, "message_tokens": 1003, "message_utf8_bytes": 2591, "cache_tensor_bytes": null, "serialized_payload_bytes": 2591, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24075.56304091122, "ttft_ms": 13.656334020197392, "decode_ms": 13.581875013187528, "receiver_prefill_ms": 117.85370798315853}, "total_wall_ms": 24222.49758290127, "ttft_ms": 13.656334020197392, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S3"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "6de94ce0a0382b5fcf2789f34d988c1249a9cec61fa53445fcb9c0d093778e0b", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "4b00ee6496d1f88b8afd2edd", "attempt_id": "05586e204ea1", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:22:56.394750Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 25}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 1099, "message_tokens": 1003, "message_utf8_bytes": 2591, "cache_tensor_bytes": null, "serialized_payload_bytes": 2591, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24075.56304091122, "ttft_ms": 14.04249994084239, "decode_ms": 13.918000040575862, "receiver_prefill_ms": 127.61645799037069}, "total_wall_ms": 24232.921874849126, "ttft_ms": 14.04249994084239, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S3"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "6de94ce0a0382b5fcf2789f34d988c1249a9cec61fa53445fcb9c0d093778e0b", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "2fdac95b6ee0928ac69fe964", "attempt_id": "73adcf3f45ad", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:23:20.751200Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 2019, "message_tokens": 1928, "message_utf8_bytes": 4983, "cache_tensor_bytes": null, "serialized_payload_bytes": 4983, "side_information_bytes": 0, "fetched_bytes": 2615, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24184.313874924555, "ttft_ms": 13.559416984207928, "decode_ms": 13.682666933164, "receiver_prefill_ms": 349.13591702934355}, "total_wall_ms": 24566.211582976393, "ttft_ms": 13.559416984207928, "input_usage": {"lookups": 9}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S3"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "52ea16eb0bfb704a11dbae8feb7b76f3e4814cc327e0d17b2df72d5e7c17ffd3", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a7fe78d98ced11deaaca9327", "attempt_id": "d1448a9b5b72", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:23:21.148134Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 1925, "message_tokens": 1834, "message_utf8_bytes": 4833, "cache_tensor_bytes": null, "serialized_payload_bytes": 4833, "side_information_bytes": 0, "fetched_bytes": 2473, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24184.313874924555, "ttft_ms": 14.246416045352817, "decode_ms": 139.0932920621708, "receiver_prefill_ms": 373.53750003967434}, "total_wall_ms": 24713.946666917764, "ttft_ms": 14.246416045352817, "input_usage": {"lookups": 9}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "HM-38K-5533", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S3"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "6571afc1c15f8522cbac9d4fded3df1343d59d06c37dd731a5a5a5f4530f26db", "output_hash": "18acd9807857243bbded3bceae55398316a3958b81d70364441bdb6748eaf878"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "4411fbc5f983c09e0e2f07b5", "attempt_id": "41b4c1c58891", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:23:21.692982Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 2035, "message_tokens": 1935, "message_utf8_bytes": 4985, "cache_tensor_bytes": null, "serialized_payload_bytes": 4985, "side_information_bytes": 0, "fetched_bytes": 2611, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24184.313874924555, "ttft_ms": 13.558916980400681, "decode_ms": 13.487166026607156, "receiver_prefill_ms": 212.6613330328837}, "total_wall_ms": 24427.022665855475, "ttft_ms": 13.558916980400681, "input_usage": {"lookups": 9}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S3"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "8a4121b6f47760dd1c31a456ccb215bf47bed6c43a9fba397bae259ff7a4fed3", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "ac07c35993d352bb65ac302d", "attempt_id": "6ba8a1dc0571", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:23:21.956238Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 2028, "message_tokens": 1935, "message_utf8_bytes": 4985, "cache_tensor_bytes": null, "serialized_payload_bytes": 4985, "side_information_bytes": 0, "fetched_bytes": 2611, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24184.313874924555, "ttft_ms": 13.441957999020815, "decode_ms": 93.9135419903323, "receiver_prefill_ms": 211.67083305772394}, "total_wall_ms": 24506.164749967866, "ttft_ms": 13.441957999020815, "input_usage": {"lookups": 9}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8340685", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S3"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "e03d1ece86798f5b635077a1595a0e43cadf471beb55d2c14af11242769a5172", "output_hash": "760ee2cb6af672573a501727b8df5f0904f8acd34231bf3af02ca9fd3fb0ff55"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "6847bab39e8ae47b93769f1d", "attempt_id": "24d33521450b", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:23:22.292826Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 1961, "message_tokens": 1860, "message_utf8_bytes": 4855, "cache_tensor_bytes": null, "serialized_payload_bytes": 4855, "side_information_bytes": 0, "fetched_bytes": 2499, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24184.313874924555, "ttft_ms": 13.964042067527771, "decode_ms": 83.87358405161649, "receiver_prefill_ms": 350.6950420560315}, "total_wall_ms": 24635.621416848153, "ttft_ms": 13.964042067527771, "input_usage": {"lookups": 9}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Melymi-15", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S3"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "b9567118acb31d8f31d64d9920c14dffe4ab19a55af3d018de7a5ad9abb22b37", "output_hash": "fa08d1474b01d6dfbf3dea049be251aa3c3021c17b4818258064b156509af58b"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b3f4e67015b85cc12b821ec0", "attempt_id": "520d749edaba", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:23:22.758176Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 2028, "message_tokens": 1935, "message_utf8_bytes": 4985, "cache_tensor_bytes": null, "serialized_payload_bytes": 4985, "side_information_bytes": 0, "fetched_bytes": 2611, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24184.313874924555, "ttft_ms": 13.70666699949652, "decode_ms": 94.84700008761138, "receiver_prefill_ms": 212.99783303402364}, "total_wall_ms": 24508.685124921612, "ttft_ms": 13.70666699949652, "input_usage": {"lookups": 9}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1346526", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S3"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "4806019e9427ed07d85b6352aa79ced77ea9085fa50b1569a548edd049a3dea6", "output_hash": "cae69acecb66385ff53349a3ef3a35e8f91348151d05f29cafeb84cf99f3fcc1"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "3713512d3329f01feba92241", "attempt_id": "62154807acbc", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:23:23.097359Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 1961, "message_tokens": 1867, "message_utf8_bytes": 4907, "cache_tensor_bytes": null, "serialized_payload_bytes": 4907, "side_information_bytes": 0, "fetched_bytes": 2537, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24184.313874924555, "ttft_ms": 14.017459005117416, "decode_ms": 13.609625049866736, "receiver_prefill_ms": 217.37566695082933}, "total_wall_ms": 24432.09533393383, "ttft_ms": 14.017459005117416, "input_usage": {"lookups": 9}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S3"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "5f3f8f8a9fc3afd11d1ad5d728f023f76f8a00c876c5b5e02555af5bf839ee6b", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "86e505223529967a1aaf489a", "attempt_id": "1a390b4b85c6", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:23:23.359773Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 2009, "message_tokens": 1914, "message_utf8_bytes": 4940, "cache_tensor_bytes": null, "serialized_payload_bytes": 4940, "side_information_bytes": 0, "fetched_bytes": 2577, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24184.313874924555, "ttft_ms": 13.744291965849698, "decode_ms": 13.493375037796795, "receiver_prefill_ms": 231.32062505465}, "total_wall_ms": 24445.693041896448, "ttft_ms": 13.744291965849698, "input_usage": {"lookups": 9}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S3"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a5fff3f9baaacb9a4a6d44f1a7a0c0b50e638151aa066a24065f067f7e40290e", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "7928945696f38b9bcaa18668", "attempt_id": "fb50eed85c28", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:23:24.497686Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 94003200}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 45, "message_tokens": null, "message_utf8_bytes": 8663, "cache_tensor_bytes": 94003200, "serialized_payload_bytes": 94003200, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 861.7216250859201, "ttft_ms": 120.53958303295076, "decode_ms": 84.57408309914172, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1078.6046250723302, "ttft_ms": 120.53958303295076, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "272184", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S3"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "3aadaddf09c30b8bfe5dc7e2aa6984fcf0df4d216145cf925760dd5a946d6e05"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "3ba290fcfb1b7db431f12e2f", "attempt_id": "ffb0406e8fa3", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:23:24.731126Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 94003200}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 45, "message_tokens": null, "message_utf8_bytes": 8663, "cache_tensor_bytes": 94003200, "serialized_payload_bytes": 94003200, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 861.7216250859201, "ttft_ms": 28.920709039084613, "decode_ms": 139.82287503313273, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1032.6261250302196, "ttft_ms": 28.920709039084613, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "HQ-34D-9894", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S3"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "0b272ad8b13ce9a89c2f8518b664e9bde3db29f866a989729649ee9e9ff6b1b4"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "317e888f5bebc2ce900eea33", "attempt_id": "0627a221a0e2", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:23:24.916280Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 94003200}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 54, "message_tokens": null, "message_utf8_bytes": 8663, "cache_tensor_bytes": 94003200, "serialized_payload_bytes": 94003200, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 861.7216250859201, "ttft_ms": 96.80733305867761, "decode_ms": 126.97820807807148, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1088.4034170303494, "ttft_ms": 96.80733305867761, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "3,237,105", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S3"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "2b6e34c5cab8756937b56b521d3c0a04fbaf070c16770171c09907f3e561dcc1"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "6fd6ecd138aa86fb9b3cb652", "attempt_id": "9498cd1def06", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:23:25.156775Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 94003200}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 8663, "cache_tensor_bytes": 94003200, "serialized_payload_bytes": 94003200, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 861.7216250859201, "ttft_ms": 29.801500029861927, "decode_ms": 125.8689999813214, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1020.5234590684995, "ttft_ms": 29.801500029861927, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "3,237,105", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S3"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "2b6e34c5cab8756937b56b521d3c0a04fbaf070c16770171c09907f3e561dcc1"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a3600089462f1c9820ced325", "attempt_id": "6299f6dd19b5", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:23:25.329944Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 94003200}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 55, "message_tokens": null, "message_utf8_bytes": 8663, "cache_tensor_bytes": 94003200, "serialized_payload_bytes": 94003200, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 861.7216250859201, "ttft_ms": 57.243875111453235, "decode_ms": 83.8574580848217, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1005.8798751560971, "ttft_ms": 57.243875111453235, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Ruknov-12", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S3"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "f7aae1e5b23999fd5f007dabe7ac59df11265b20b68c84615515729028871151"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "ac498076e109b255b838d6ed", "attempt_id": "b5aca9e2274d", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:23:25.488516Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 94003200}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 8663, "cache_tensor_bytes": 94003200, "serialized_payload_bytes": 94003200, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 861.7216250859201, "ttft_ms": 29.81895802076906, "decode_ms": 97.42037497926503, "receiver_prefill_ms": 0.0}, "total_wall_ms": 992.1437910525128, "ttft_ms": 29.81895802076906, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "6049819", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S3"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "bd56e0b4883215bd629f92f9caa788b9983bd41635f1f197e2663d34e52edd7f"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "e7512d9f64e9b0a3398d4e83", "attempt_id": "7c400e71d975", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:23:25.633088Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 94003200}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 48, "message_tokens": null, "message_utf8_bytes": 8663, "cache_tensor_bytes": 94003200, "serialized_payload_bytes": 94003200, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 861.7216250859201, "ttft_ms": 59.02858299668878, "decode_ms": 14.062582980841398, "receiver_prefill_ms": 0.0}, "total_wall_ms": 937.7773331943899, "ttft_ms": 59.02858299668878, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S3"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "7842eed94911a656fff6af57", "attempt_id": "b52b45a24664", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:23:25.723448Z", "task_family": "synthetic", "document_id": "syn_03018_2000", "document_hash": "d284eb07e27da742b9c2b67abccfdf72e4dd67c3259f44f3b41e34dd3e6fc1ff", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03018_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 94003200}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1982, "query_tokens": 49, "message_tokens": null, "message_utf8_bytes": 8663, "cache_tensor_bytes": 94003200, "serialized_payload_bytes": 94003200, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 861.7216250859201, "ttft_ms": 30.05533292889595, "decode_ms": 13.99841601960361, "receiver_prefill_ms": 0.0}, "total_wall_ms": 908.8780410820618, "ttft_ms": 30.05533292889595, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Eshtorlun-12", "bogus_amount": "283,475", "source": "S3"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03018_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c90a286f101d03469af7dc29", "attempt_id": "ce03ebd0aa83", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:23:26.749574Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 6187, "message_tokens": 6103, "message_utf8_bytes": 30078, "cache_tensor_bytes": null, "serialized_payload_bytes": 30078, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.03758410457521677, "ttft_ms": 13.891124981455505, "decode_ms": 98.23600004892796, "receiver_prefill_ms": 740.1176249841228}, "total_wall_ms": 863.9176260912791, "ttft_ms": 13.891124981455505, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8544550", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "f954bc741458afc161d3b4a4fb4f7fe52425f5f508bb1d60133554fadf04a6fd", "output_hash": "7ea69da7ca13297bce5fac6a79b7f74d78548d728dc85a91c943d1c2c8172768"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "3bfe27a26779ce19f594bec1", "attempt_id": "1f845244c402", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:23:27.630435Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 6189, "message_tokens": 6103, "message_utf8_bytes": 30078, "cache_tensor_bytes": null, "serialized_payload_bytes": 30078, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.03758410457521677, "ttft_ms": 14.184083906002343, "decode_ms": 153.4824169939384, "receiver_prefill_ms": 723.0160420294851}, "total_wall_ms": 900.3609180217609, "ttft_ms": 14.184083906002343, "input_usage": {}, "output_usage": {"output_tokens": 11}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "JX-64S-4555", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "f954bc741458afc161d3b4a4fb4f7fe52425f5f508bb1d60133554fadf04a6fd", "output_hash": "f6f8ca3da45dc6c01a3dede99eab98f0f3fd2d2c5e105791d9c68c607877ed22"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b03db8c230d582c85d299eda", "attempt_id": "b27fa85d9d16", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:23:28.547567Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 6195, "message_tokens": 6103, "message_utf8_bytes": 30078, "cache_tensor_bytes": null, "serialized_payload_bytes": 30078, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.03758410457521677, "ttft_ms": 13.973666005767882, "decode_ms": 98.60866703093052, "receiver_prefill_ms": 722.3557910183445}, "total_wall_ms": 844.4480841280892, "ttft_ms": 13.973666005767882, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8859765", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "f954bc741458afc161d3b4a4fb4f7fe52425f5f508bb1d60133554fadf04a6fd", "output_hash": "8d800485162553e438ca8ddff7de5613a4800ab0aeecee4f411e1cec8e2008bd"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "33e419a36fc3b9901bd135ab", "attempt_id": "a3348a9cf0cb", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:23:29.408628Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 6190, "message_tokens": 6103, "message_utf8_bytes": 30078, "cache_tensor_bytes": null, "serialized_payload_bytes": 30078, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.03758410457521677, "ttft_ms": 13.830917072482407, "decode_ms": 97.2098329802975, "receiver_prefill_ms": 723.0467500630766}, "total_wall_ms": 843.7343341065571, "ttft_ms": 13.830917072482407, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7730705", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "f954bc741458afc161d3b4a4fb4f7fe52425f5f508bb1d60133554fadf04a6fd", "output_hash": "6c885473fb45b44c6a425cd3cd307b4771e35351a0e2c17186a92d2de9c1fd42"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "dbcc69e60cc946c1c4925859", "attempt_id": "909282747386", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "d40b392c23efff803ea24b0c3a963a6e8aab4a32", "timestamp_utc": "2026-10-30T08:23:30.268946Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 6198, "message_tokens": 6103, "message_utf8_bytes": 30078, "cache_tensor_bytes": null, "serialized_payload_bytes": 30078, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.03758410457521677, "ttft_ms": 13.933582929894328, "decode_ms": 69.48795798234642, "receiver_prefill_ms": 723.6029580235481}, "total_wall_ms": 816.8121670605615, "ttft_ms": 13.933582929894328, "input_usage": {}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Ixbanruk-3", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "f954bc741458afc161d3b4a4fb4f7fe52425f5f508bb1d60133554fadf04a6fd", "output_hash": "1cc74509d5485de66115e1f6b591a4a84f89826fbba3b26cfbf7cfcfdc216bf5"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "88acb85ee9df5a9f985a5548", "attempt_id": "082b381aa6f4", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:23:31.102452Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 6192, "message_tokens": 6103, "message_utf8_bytes": 30078, "cache_tensor_bytes": null, "serialized_payload_bytes": 30078, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.03758410457521677, "ttft_ms": 13.669125037267804, "decode_ms": 137.16050004586577, "receiver_prefill_ms": 723.7758749397472}, "total_wall_ms": 884.1153340181336, "ttft_ms": 13.669125037267804, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "20,478,689", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "f954bc741458afc161d3b4a4fb4f7fe52425f5f508bb1d60133554fadf04a6fd", "output_hash": "7b80aa4c363cfa59614b01d8fefd3b391f7a82483164746113ad4f4da7e7497f"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "9bc6680b620c8ca1bfd6d7d1", "attempt_id": "c0fd4e62409d", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:23:32.003243Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 6193, "message_tokens": 6103, "message_utf8_bytes": 30078, "cache_tensor_bytes": null, "serialized_payload_bytes": 30078, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.03758410457521677, "ttft_ms": 13.773542013950646, "decode_ms": 13.767624972388148, "receiver_prefill_ms": 735.2882920531556}, "total_wall_ms": 772.2675011027604, "ttft_ms": 13.773542013950646, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "f954bc741458afc161d3b4a4fb4f7fe52425f5f508bb1d60133554fadf04a6fd", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "249919015871075346a00791", "attempt_id": "dddb0fa11a8c", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:23:32.792822Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 6194, "message_tokens": 6103, "message_utf8_bytes": 30078, "cache_tensor_bytes": null, "serialized_payload_bytes": 30078, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.03758410457521677, "ttft_ms": 14.056417043320835, "decode_ms": 14.182667015120387, "receiver_prefill_ms": 749.0845830179751}, "total_wall_ms": 786.9983341079205, "ttft_ms": 14.056417043320835, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "f954bc741458afc161d3b4a4fb4f7fe52425f5f508bb1d60133554fadf04a6fd", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "fd1d66941740eeaf2778b092", "attempt_id": "b237860d0d9d", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:23:48.187681Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 1123, "message_tokens": 1037, "message_utf8_bytes": 2613, "cache_tensor_bytes": null, "serialized_payload_bytes": 2613, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14590.413291938603, "ttft_ms": 13.471833080984652, "decode_ms": 120.94187503680587, "receiver_prefill_ms": 121.619459008798}, "total_wall_ms": 14853.483874932863, "ttft_ms": 13.471833080984652, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8,544,550", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "9eb44ba658f639bdbaf1826fa289b0eb2817b44d395237da065f5be90c480044", "output_hash": "275c8abab9d6375639c672e5391f7a9637810a7c2d3985510d6d956c3cb3a509"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "f7200a0ddfc418bed4fd88f7", "attempt_id": "101248486af3", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:23:48.469133Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 1125, "message_tokens": 1037, "message_utf8_bytes": 2613, "cache_tensor_bytes": null, "serialized_payload_bytes": 2613, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14590.413291938603, "ttft_ms": 13.41891591437161, "decode_ms": 146.8593750614673, "receiver_prefill_ms": 120.86083402391523}, "total_wall_ms": 14873.448999947868, "ttft_ms": 13.41891591437161, "input_usage": {}, "output_usage": {"output_tokens": 11}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "JX-64S-4555", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "9eb44ba658f639bdbaf1826fa289b0eb2817b44d395237da065f5be90c480044", "output_hash": "f6f8ca3da45dc6c01a3dede99eab98f0f3fd2d2c5e105791d9c68c607877ed22"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "e13eec89371d8d87d13805eb", "attempt_id": "0597df357773", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:23:48.770133Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 1131, "message_tokens": 1037, "message_utf8_bytes": 2613, "cache_tensor_bytes": null, "serialized_payload_bytes": 2613, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14590.413291938603, "ttft_ms": 13.578584068454802, "decode_ms": 93.96445902530104, "receiver_prefill_ms": 121.40875007025898}, "total_wall_ms": 14821.288750972599, "ttft_ms": 13.578584068454802, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8859765", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "9eb44ba658f639bdbaf1826fa289b0eb2817b44d395237da065f5be90c480044", "output_hash": "8d800485162553e438ca8ddff7de5613a4800ab0aeecee4f411e1cec8e2008bd"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "0c18c6e728c6386a37e57be8", "attempt_id": "8fe01769ff13", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:23:49.018004Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 1126, "message_tokens": 1037, "message_utf8_bytes": 2613, "cache_tensor_bytes": null, "serialized_payload_bytes": 2613, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14590.413291938603, "ttft_ms": 13.435499975457788, "decode_ms": 93.9098340459168, "receiver_prefill_ms": 121.61587504670024}, "total_wall_ms": 14821.260124910623, "ttft_ms": 13.435499975457788, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "9781740", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "9eb44ba658f639bdbaf1826fa289b0eb2817b44d395237da065f5be90c480044", "output_hash": "8b57ccd1ed84757887221612d34399545b6be9ebd8480805f7c2b5ffaa4dea1c"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "1403e7d54e48b4591b1d0e4b", "attempt_id": "4e0dfe60abff", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:23:49.266277Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 1134, "message_tokens": 1037, "message_utf8_bytes": 2613, "cache_tensor_bytes": null, "serialized_payload_bytes": 2613, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14590.413291938603, "ttft_ms": 13.573667034506798, "decode_ms": 67.13166704867035, "receiver_prefill_ms": 121.31224991753697}, "total_wall_ms": 14794.301833957434, "ttft_ms": 13.573667034506798, "input_usage": {}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Ixbanruk-3", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "9eb44ba658f639bdbaf1826fa289b0eb2817b44d395237da065f5be90c480044", "output_hash": "1cc74509d5485de66115e1f6b591a4a84f89826fbba3b26cfbf7cfcfdc216bf5"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "275befd3f80b78ad0a685ec9", "attempt_id": "cdc0cb1ad4dd", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:23:49.487660Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 1128, "message_tokens": 1037, "message_utf8_bytes": 2613, "cache_tensor_bytes": null, "serialized_payload_bytes": 2613, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14590.413291938603, "ttft_ms": 13.37075000628829, "decode_ms": 107.46566706802696, "receiver_prefill_ms": 121.50362506508827}, "total_wall_ms": 14834.61320900824, "ttft_ms": 13.37075000628829, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "20458969", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "9eb44ba658f639bdbaf1826fa289b0eb2817b44d395237da065f5be90c480044", "output_hash": "7756e3d6c27c609ae9b5d5d8e4846e6ce5ce0ce61feccfe4eca30350985c4dd7"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "fe28b7299176a54d413f13f0", "attempt_id": "3d828d1f0034", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:23:49.749143Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 1129, "message_tokens": 1037, "message_utf8_bytes": 2613, "cache_tensor_bytes": null, "serialized_payload_bytes": 2613, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14590.413291938603, "ttft_ms": 13.388750026933849, "decode_ms": 13.4045829763636, "receiver_prefill_ms": 121.25162500888109}, "total_wall_ms": 14740.30895892065, "ttft_ms": 13.388750026933849, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "9eb44ba658f639bdbaf1826fa289b0eb2817b44d395237da065f5be90c480044", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "f5ea8c2e92bd225fe988f7f4", "attempt_id": "5a9a72c9d8b1", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:23:49.915803Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 1130, "message_tokens": 1037, "message_utf8_bytes": 2613, "cache_tensor_bytes": null, "serialized_payload_bytes": 2613, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14590.413291938603, "ttft_ms": 13.584292028099298, "decode_ms": 13.418916962109506, "receiver_prefill_ms": 121.12466595135629}, "total_wall_ms": 14740.360624855384, "ttft_ms": 13.584292028099298, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "9eb44ba658f639bdbaf1826fa289b0eb2817b44d395237da065f5be90c480044", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a4e401a3a2e19b15a8d6600e", "attempt_id": "abb785bc653a", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:24:02.912395Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 15}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 659, "message_tokens": 570, "message_utf8_bytes": 1631, "cache_tensor_bytes": null, "serialized_payload_bytes": 1631, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 12830.007417011075, "ttft_ms": 13.909582979977131, "decode_ms": 136.0027079936117, "receiver_prefill_ms": 75.20962494891137}, "total_wall_ms": 13056.573792011477, "ttft_ms": 13.909582979977131, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5,926,224 CAD", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "3ed2d5b7c9724067efc8c50c7e6e06d122e8f30ce4a7f22dfc861b7b3481da52", "output_hash": "469758eb51aa9aa28367572d59bfb330d2224ef4b69a21eca47c0980dacbcc3b"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b4d8e56ded3852ed6272072a", "attempt_id": "ea9f13f52b30", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:24:03.157396Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 15}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 661, "message_tokens": 570, "message_utf8_bytes": 1631, "cache_tensor_bytes": null, "serialized_payload_bytes": 1631, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 12830.007417011075, "ttft_ms": 13.384709018282592, "decode_ms": 147.6515829563141, "receiver_prefill_ms": 75.23533294443041}, "total_wall_ms": 13067.65662599355, "ttft_ms": 13.384709018282592, "input_usage": {}, "output_usage": {"output_tokens": 11}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "JX-64S-4555", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "3ed2d5b7c9724067efc8c50c7e6e06d122e8f30ce4a7f22dfc861b7b3481da52", "output_hash": "f6f8ca3da45dc6c01a3dede99eab98f0f3fd2d2c5e105791d9c68c607877ed22"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "7feb10fe60a8068ef943ed44", "attempt_id": "f7f1cb0b7e7d", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:24:03.413974Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 15}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 667, "message_tokens": 570, "message_utf8_bytes": 1631, "cache_tensor_bytes": null, "serialized_payload_bytes": 1631, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 12830.007417011075, "ttft_ms": 13.376667047850788, "decode_ms": 107.30066697578877, "receiver_prefill_ms": 75.11304109357297}, "total_wall_ms": 13027.133292052895, "ttft_ms": 13.376667047850788, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "9119.650", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "3ed2d5b7c9724067efc8c50c7e6e06d122e8f30ce4a7f22dfc861b7b3481da52", "output_hash": "a758b714660e7268fc27a5814beb28446c47a2ebb89e12e43c736cd5276a4e45"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a5d831169b3d00fb5af20e7c", "attempt_id": "2ded618c2416", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:24:03.629071Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 15}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 662, "message_tokens": 570, "message_utf8_bytes": 1631, "cache_tensor_bytes": null, "serialized_payload_bytes": 1631, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 12830.007417011075, "ttft_ms": 13.329041074030101, "decode_ms": 13.33454204723239, "receiver_prefill_ms": 74.92666598409414}, "total_wall_ms": 12932.93000094127, "ttft_ms": 13.329041074030101, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "3ed2d5b7c9724067efc8c50c7e6e06d122e8f30ce4a7f22dfc861b7b3481da52", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "9644cfcf7a4887192c0cfe01", "attempt_id": "dc6939d16f51", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:24:03.748879Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 15}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 670, "message_tokens": 570, "message_utf8_bytes": 1631, "cache_tensor_bytes": null, "serialized_payload_bytes": 1631, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 12830.007417011075, "ttft_ms": 13.60695797484368, "decode_ms": 53.31479199230671, "receiver_prefill_ms": 75.0788749428466}, "total_wall_ms": 12973.385166958906, "ttft_ms": 13.60695797484368, "input_usage": {}, "output_usage": {"output_tokens": 4}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Melruk-3", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "3ed2d5b7c9724067efc8c50c7e6e06d122e8f30ce4a7f22dfc861b7b3481da52", "output_hash": "29e02c51d1e6eeabd8d3fcec3b482c44c02ebc350a5d4b73304fcba97294db93"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "51a4948ae15fc72148de8458", "attempt_id": "e628d1ff5323", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:24:03.909216Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 15}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 664, "message_tokens": 570, "message_utf8_bytes": 1631, "cache_tensor_bytes": null, "serialized_payload_bytes": 1631, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 12830.007417011075, "ttft_ms": 13.44929204788059, "decode_ms": 93.3201250154525, "receiver_prefill_ms": 74.76579106878489}, "total_wall_ms": 13012.867292040028, "ttft_ms": 13.44929204788059, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5926224", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "3ed2d5b7c9724067efc8c50c7e6e06d122e8f30ce4a7f22dfc861b7b3481da52", "output_hash": "7c1113578ca4e75c3dd3a916eae945b8147de18f2e9b429a97f3e3481fa7d2fc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "7b5393a901d5c11b10da5201", "attempt_id": "e365416a9c60", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:24:04.109336Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 15}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 665, "message_tokens": 570, "message_utf8_bytes": 1631, "cache_tensor_bytes": null, "serialized_payload_bytes": 1631, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 12830.007417011075, "ttft_ms": 13.426417019218206, "decode_ms": 13.1738749332726, "receiver_prefill_ms": 74.56199999433011}, "total_wall_ms": 12932.435624999925, "ttft_ms": 13.426417019218206, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "3ed2d5b7c9724067efc8c50c7e6e06d122e8f30ce4a7f22dfc861b7b3481da52", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "9049976a51c8218dab3c5a98", "attempt_id": "e38bdfa7c949", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:24:04.229236Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 15}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 666, "message_tokens": 570, "message_utf8_bytes": 1631, "cache_tensor_bytes": null, "serialized_payload_bytes": 1631, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 12830.007417011075, "ttft_ms": 13.150707934983075, "decode_ms": 13.477290980517864, "receiver_prefill_ms": 74.24574997276068}, "total_wall_ms": 12932.265999959782, "ttft_ms": 13.150707934983075, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "3ed2d5b7c9724067efc8c50c7e6e06d122e8f30ce4a7f22dfc861b7b3481da52", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "bbf707c30968d770eef51d16", "attempt_id": "51061678de67", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:24:17.152260Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 714, "message_tokens": 626, "message_utf8_bytes": 1629, "cache_tensor_bytes": null, "serialized_payload_bytes": 1629, "side_information_bytes": 0, "fetched_bytes": 926, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 12803.8252920378, "ttft_ms": 13.277083984576166, "decode_ms": 93.5497919563204, "receiver_prefill_ms": 78.69395799934864}, "total_wall_ms": 12991.238376125693, "ttft_ms": 13.277083984576166, "input_usage": {"lookups": 3}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8544550", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "784f9df6e6021c164bc4f5667279585d1722c0ab1ac935b2028a0cd6485dd48a", "output_hash": "7ea69da7ca13297bce5fac6a79b7f74d78548d728dc85a91c943d1c2c8172768"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "363c4e7ca56070ddc327624f", "attempt_id": "a53a3a39a72a", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:24:17.362412Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 716, "message_tokens": 626, "message_utf8_bytes": 1629, "cache_tensor_bytes": null, "serialized_payload_bytes": 1629, "side_information_bytes": 0, "fetched_bytes": 926, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 12803.8252920378, "ttft_ms": 13.364790938794613, "decode_ms": 13.386250007897615, "receiver_prefill_ms": 77.86524994298816}, "total_wall_ms": 12909.889208967797, "ttft_ms": 13.364790938794613, "input_usage": {"lookups": 3}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "784f9df6e6021c164bc4f5667279585d1722c0ab1ac935b2028a0cd6485dd48a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "133b20460c7b16b478d5611b", "attempt_id": "af231495fbbb", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:24:17.485275Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 687, "message_tokens": 591, "message_utf8_bytes": 1619, "cache_tensor_bytes": null, "serialized_payload_bytes": 1619, "side_information_bytes": 0, "fetched_bytes": 920, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 12803.8252920378, "ttft_ms": 13.611791073344648, "decode_ms": 92.91875001508743, "receiver_prefill_ms": 76.76341698970646}, "total_wall_ms": 12988.52620809339, "ttft_ms": 13.611791073344648, "input_usage": {"lookups": 3}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8995865", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "16bef4f5181f17888ee076c3812965f375bb03049319c8e2b4e112e86524ff97", "output_hash": "bce25ee55f2656e972d49f9c0a22f2219979f1bd4e5965f2424a502257eef0a9"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "51dcb2487ea2265bf6871b6a", "attempt_id": "8082169b51ce", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:24:17.687105Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 717, "message_tokens": 626, "message_utf8_bytes": 1629, "cache_tensor_bytes": null, "serialized_payload_bytes": 1629, "side_information_bytes": 0, "fetched_bytes": 926, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 12803.8252920378, "ttft_ms": 13.30454193521291, "decode_ms": 93.2469159597531, "receiver_prefill_ms": 77.68229197245091}, "total_wall_ms": 12989.51745906379, "ttft_ms": 13.30454193521291, "input_usage": {"lookups": 3}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "9781740", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "c1e6f711d2e4793023f695e01e3d4e6522e46db8793e480a347dc408c6191258", "output_hash": "8b57ccd1ed84757887221612d34399545b6be9ebd8480805f7c2b5ffaa4dea1c"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "cd8be5aa65eca31717f4f69b", "attempt_id": "99554637cfac", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:24:17.890085Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 725, "message_tokens": 626, "message_utf8_bytes": 1629, "cache_tensor_bytes": null, "serialized_payload_bytes": 1629, "side_information_bytes": 0, "fetched_bytes": 926, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 12803.8252920378, "ttft_ms": 13.529415940865874, "decode_ms": 66.58237497322261, "receiver_prefill_ms": 78.45983398146927}, "total_wall_ms": 12963.804542087018, "ttft_ms": 13.529415940865874, "input_usage": {"lookups": 3}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Ixbanruk-3", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "784f9df6e6021c164bc4f5667279585d1722c0ab1ac935b2028a0cd6485dd48a", "output_hash": "1cc74509d5485de66115e1f6b591a4a84f89826fbba3b26cfbf7cfcfdc216bf5"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "bd86d6e72f8f782566602058", "attempt_id": "3bcc6ab9d77b", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:24:18.067430Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 719, "message_tokens": 626, "message_utf8_bytes": 1629, "cache_tensor_bytes": null, "serialized_payload_bytes": 1629, "side_information_bytes": 0, "fetched_bytes": 926, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 12803.8252920378, "ttft_ms": 13.115542009472847, "decode_ms": 132.8475830378011, "receiver_prefill_ms": 78.30358401406556}, "total_wall_ms": 13029.5239581028, "ttft_ms": 13.115542009472847, "input_usage": {"lookups": 3}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "32,261,765", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "784f9df6e6021c164bc4f5667279585d1722c0ab1ac935b2028a0cd6485dd48a", "output_hash": "970de185b6e73d597ba9e045394d564cfc69d8e36403c3ab2c02ee74da377ad9"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "e85d9f6f6fd17020f4b24f31", "attempt_id": "7619d48e2587", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:24:18.310056Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 720, "message_tokens": 626, "message_utf8_bytes": 1629, "cache_tensor_bytes": null, "serialized_payload_bytes": 1629, "side_information_bytes": 0, "fetched_bytes": 926, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 12803.8252920378, "ttft_ms": 13.520332984626293, "decode_ms": 13.20024998858571, "receiver_prefill_ms": 77.8793750796467}, "total_wall_ms": 12909.82833399903, "ttft_ms": 13.520332984626293, "input_usage": {"lookups": 3}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "784f9df6e6021c164bc4f5667279585d1722c0ab1ac935b2028a0cd6485dd48a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "0799fbeff837fb36192118fa", "attempt_id": "c69dc620449e", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:24:18.433263Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 721, "message_tokens": 626, "message_utf8_bytes": 1629, "cache_tensor_bytes": null, "serialized_payload_bytes": 1629, "side_information_bytes": 0, "fetched_bytes": 926, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 12803.8252920378, "ttft_ms": 13.208749936893582, "decode_ms": 13.199875014834106, "receiver_prefill_ms": 78.20004201494157}, "total_wall_ms": 12909.903916995972, "ttft_ms": 13.208749936893582, "input_usage": {"lookups": 3}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "784f9df6e6021c164bc4f5667279585d1722c0ab1ac935b2028a0cd6485dd48a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b8e915a084ce517a4039bf43", "attempt_id": "e457b533a060", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:24:19.291547Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 227856384}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 42, "message_tokens": null, "message_utf8_bytes": 30052, "cache_tensor_bytes": 227856384, "serialized_payload_bytes": 227856384, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 735.037749982439, "ttft_ms": 53.70887496974319, "decode_ms": 106.23174998909235, "receiver_prefill_ms": 0.0}, "total_wall_ms": 904.3665840290487, "ttft_ms": 53.70887496974319, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8544550", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "7ea69da7ca13297bce5fac6a79b7f74d78548d728dc85a91c943d1c2c8172768"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c22b5eff84682d6098bf0f13", "attempt_id": "9b7cd5663932", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:24:19.478226Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 227856384}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 44, "message_tokens": null, "message_utf8_bytes": 30052, "cache_tensor_bytes": 227856384, "serialized_payload_bytes": 227856384, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 735.037749982439, "ttft_ms": 48.04024996701628, "decode_ms": 165.4460000572726, "receiver_prefill_ms": 0.0}, "total_wall_ms": 955.4720829473808, "ttft_ms": 48.04024996701628, "input_usage": {}, "output_usage": {"output_tokens": 11}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "JX-67G-1446", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "17765d5cd1a65fba6a352dc360b5880164ca5fbfb993c9932800bca1d99b7480"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "bb7df9f6acbdba0bf171a05d", "attempt_id": "3aeea65721e9", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:24:19.715831Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 227856384}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 50, "message_tokens": null, "message_utf8_bytes": 30052, "cache_tensor_bytes": 227856384, "serialized_payload_bytes": 227856384, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 735.037749982439, "ttft_ms": 48.31158299930394, "decode_ms": 104.78958394378424, "receiver_prefill_ms": 0.0}, "total_wall_ms": 895.5197499599308, "ttft_ms": 48.31158299930394, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8995865", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "bce25ee55f2656e972d49f9c0a22f2219979f1bd4e5965f2424a502257eef0a9"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "45acd3251e8c45aca30693ff", "attempt_id": "95e754255487", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:24:19.893885Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 227856384}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 45, "message_tokens": null, "message_utf8_bytes": 30052, "cache_tensor_bytes": 227856384, "serialized_payload_bytes": 227856384, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 735.037749982439, "ttft_ms": 47.06091689877212, "decode_ms": 105.19037500489503, "receiver_prefill_ms": 0.0}, "total_wall_ms": 894.6441250154749, "ttft_ms": 47.06091689877212, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7730705", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "6c885473fb45b44c6a425cd3cd307b4771e35351a0e2c17186a92d2de9c1fd42"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b0a7854308b71e8b7cde5901", "attempt_id": "be4d11672b5c", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:24:20.070662Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 227856384}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 53, "message_tokens": null, "message_utf8_bytes": 30052, "cache_tensor_bytes": 227856384, "serialized_payload_bytes": 227856384, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 735.037749982439, "ttft_ms": 48.55720803607255, "decode_ms": 74.8714170185849, "receiver_prefill_ms": 0.0}, "total_wall_ms": 865.6046659452841, "ttft_ms": 48.55720803607255, "input_usage": {}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Ixbanruk-3", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "1cc74509d5485de66115e1f6b591a4a84f89826fbba3b26cfbf7cfcfdc216bf5"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "6ee3ce4cf1af3be44d3abfba", "attempt_id": "d772efc29f40", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:24:20.218488Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 227856384}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 30052, "cache_tensor_bytes": 227856384, "serialized_payload_bytes": 227856384, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 735.037749982439, "ttft_ms": 45.48958293162286, "decode_ms": 149.7669160598889, "receiver_prefill_ms": 0.0}, "total_wall_ms": 937.8280839882791, "ttft_ms": 45.48958293162286, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "20,647,154", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "0707b34b969871d1901b99df68dc3821da0c02077f67e1c7c6030289924e4363"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "3a2b4093ffd01cd7a8aefb70", "attempt_id": "0b219002f0a1", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:24:20.438636Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 227856384}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 48, "message_tokens": null, "message_utf8_bytes": 30052, "cache_tensor_bytes": 227856384, "serialized_payload_bytes": 227856384, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 735.037749982439, "ttft_ms": 47.02149995137006, "decode_ms": 15.020041028037667, "receiver_prefill_ms": 0.0}, "total_wall_ms": 804.4304579962045, "ttft_ms": 47.02149995137006, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "cc037e488023ec57146d1d5b", "attempt_id": "fa41d77846d7", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:24:20.524876Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 227856384}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 49, "message_tokens": null, "message_utf8_bytes": 30052, "cache_tensor_bytes": 227856384, "serialized_payload_bytes": 227856384, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 735.037749982439, "ttft_ms": 46.92787502426654, "decode_ms": 14.99320799484849, "receiver_prefill_ms": 0.0}, "total_wall_ms": 803.4068749984726, "ttft_ms": 46.92787502426654, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b4e6b69aaeb12a203f9ccf72", "attempt_id": "bf97bc85cceb", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:24:20.610292Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 6301, "message_tokens": 6217, "message_utf8_bytes": 30582, "cache_tensor_bytes": null, "serialized_payload_bytes": 30582, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.036457902751863, "ttft_ms": 13.901624944992363, "decode_ms": 123.75337490811944, "receiver_prefill_ms": 747.6423329208046}, "total_wall_ms": 895.0238329125568, "ttft_ms": 13.901624944992363, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8,544,550", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S4"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "b702fdc180b60aa41840fd0f3fd93ce8396baf534a968c6d07646f9bb039f461", "output_hash": "275c8abab9d6375639c672e5391f7a9637810a7c2d3985510d6d956c3cb3a509"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "32e300742405ffc74d91f530", "attempt_id": "4445e6788cc1", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:24:21.521561Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 6303, "message_tokens": 6217, "message_utf8_bytes": 30582, "cache_tensor_bytes": null, "serialized_payload_bytes": 30582, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.036457902751863, "ttft_ms": 14.233874971978366, "decode_ms": 152.04033302143216, "receiver_prefill_ms": 759.1276670573279}, "total_wall_ms": 935.2202918380499, "ttft_ms": 14.233874971978366, "input_usage": {}, "output_usage": {"output_tokens": 11}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "JX-64S-4555", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S4"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "b702fdc180b60aa41840fd0f3fd93ce8396baf534a968c6d07646f9bb039f461", "output_hash": "f6f8ca3da45dc6c01a3dede99eab98f0f3fd2d2c5e105791d9c68c607877ed22"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "0d22bcac3fa5cfffa1416d8c", "attempt_id": "7e97a71fc331", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:24:22.473928Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 6309, "message_tokens": 6217, "message_utf8_bytes": 30582, "cache_tensor_bytes": null, "serialized_payload_bytes": 30582, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.036457902751863, "ttft_ms": 14.046124997548759, "decode_ms": 96.16337495390326, "receiver_prefill_ms": 772.1485830843449}, "total_wall_ms": 892.1603738563135, "ttft_ms": 14.046124997548759, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8859765", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S4"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "b702fdc180b60aa41840fd0f3fd93ce8396baf534a968c6d07646f9bb039f461", "output_hash": "8d800485162553e438ca8ddff7de5613a4800ab0aeecee4f411e1cec8e2008bd"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "2bf9f81d93cdf14b82a6c605", "attempt_id": "7205e95d8523", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:24:23.383933Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 6304, "message_tokens": 6217, "message_utf8_bytes": 30582, "cache_tensor_bytes": null, "serialized_payload_bytes": 30582, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.036457902751863, "ttft_ms": 13.921791920438409, "decode_ms": 96.95462498348206, "receiver_prefill_ms": 804.5001659775153}, "total_wall_ms": 925.121916923672, "ttft_ms": 13.921791920438409, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7730705", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S4"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "b702fdc180b60aa41840fd0f3fd93ce8396baf534a968c6d07646f9bb039f461", "output_hash": "6c885473fb45b44c6a425cd3cd307b4771e35351a0e2c17186a92d2de9c1fd42"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "79489f79ccda836ce28ef36a", "attempt_id": "8b123f9214d4", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:24:24.326095Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 6312, "message_tokens": 6217, "message_utf8_bytes": 30582, "cache_tensor_bytes": null, "serialized_payload_bytes": 30582, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.036457902751863, "ttft_ms": 14.041124959476292, "decode_ms": 69.7017089696601, "receiver_prefill_ms": 786.4734580507502}, "total_wall_ms": 879.8140408471227, "ttft_ms": 14.041124959476292, "input_usage": {}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Ixbanruk-3", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S4"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "b702fdc180b60aa41840fd0f3fd93ce8396baf534a968c6d07646f9bb039f461", "output_hash": "1cc74509d5485de66115e1f6b591a4a84f89826fbba3b26cfbf7cfcfdc216bf5"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "026d4d11a732cc9d22f07d10", "attempt_id": "4de259a65ae8", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:24:25.222704Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 6306, "message_tokens": 6217, "message_utf8_bytes": 30582, "cache_tensor_bytes": null, "serialized_payload_bytes": 30582, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.036457902751863, "ttft_ms": 14.483500039204955, "decode_ms": 112.81841702293605, "receiver_prefill_ms": 802.0694579463452}, "total_wall_ms": 940.1143328286707, "ttft_ms": 14.483500039204955, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "21744419", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S4"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "b702fdc180b60aa41840fd0f3fd93ce8396baf534a968c6d07646f9bb039f461", "output_hash": "12e9cb5b890550a52cda5bdc057fe69b1b0e684eaf06b09da11cc7e313cdf7a7"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "47ce83c427394e8de0bc8f3b", "attempt_id": "4a5d8a425653", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:24:26.179952Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 6307, "message_tokens": 6217, "message_utf8_bytes": 30582, "cache_tensor_bytes": null, "serialized_payload_bytes": 30582, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.036457902751863, "ttft_ms": 14.126459020189941, "decode_ms": 13.839541003108025, "receiver_prefill_ms": 798.2102080713958}, "total_wall_ms": 835.9402908245102, "ttft_ms": 14.126459020189941, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S4"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "b702fdc180b60aa41840fd0f3fd93ce8396baf534a968c6d07646f9bb039f461", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "81983e8d224228574fd61803", "attempt_id": "b39662985675", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:24:27.033836Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 6308, "message_tokens": 6217, "message_utf8_bytes": 30582, "cache_tensor_bytes": null, "serialized_payload_bytes": 30582, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.036457902751863, "ttft_ms": 14.28387500345707, "decode_ms": 13.76383297611028, "receiver_prefill_ms": 814.1360419103876}, "total_wall_ms": 852.1435419097543, "ttft_ms": 14.28387500345707, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S4"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "b702fdc180b60aa41840fd0f3fd93ce8396baf534a968c6d07646f9bb039f461", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "9334118a4bd5adeed6b56945", "attempt_id": "c179718cf38e", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:24:51.283032Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1677}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 1786, "message_tokens": 1701, "message_utf8_bytes": 6315, "cache_tensor_bytes": null, "serialized_payload_bytes": 6315, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 23379.230667022057, "ttft_ms": 13.489042059518397, "decode_ms": 123.15150001086295, "receiver_prefill_ms": 186.11733301077038}, "total_wall_ms": 23707.162084057927, "ttft_ms": 13.489042059518397, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7,730,705", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S4"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "bf1dfd9b854995ec8ba51cff051d35793afdffacd18b48cff91f7b39fb8a9ccd", "output_hash": "2e0f40d8945ba6021c3c207215c3d834da5944796960b093ea104a759c7b3bd0"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "e1137d27137888bd90a0c855", "attempt_id": "7eaad7334598", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:24:51.628576Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1677}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 1788, "message_tokens": 1701, "message_utf8_bytes": 6315, "cache_tensor_bytes": null, "serialized_payload_bytes": 6315, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 23379.230667022057, "ttft_ms": 13.404750032350421, "decode_ms": 148.78950000274926, "receiver_prefill_ms": 185.85150002036244}, "total_wall_ms": 23730.235458002426, "ttft_ms": 13.404750032350421, "input_usage": {}, "output_usage": {"output_tokens": 11}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "JX-64S-4555", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S4"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "bf1dfd9b854995ec8ba51cff051d35793afdffacd18b48cff91f7b39fb8a9ccd", "output_hash": "f6f8ca3da45dc6c01a3dede99eab98f0f3fd2d2c5e105791d9c68c607877ed22"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "cd3642d65df0fc015bb9f03c", "attempt_id": "62dbc11e03d4", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:24:51.996997Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1677}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 1794, "message_tokens": 1701, "message_utf8_bytes": 6315, "cache_tensor_bytes": null, "serialized_payload_bytes": 6315, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 23379.230667022057, "ttft_ms": 14.217249932698905, "decode_ms": 94.61891697719693, "receiver_prefill_ms": 193.10587493237108}, "total_wall_ms": 23684.089625021443, "ttft_ms": 14.217249932698905, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8859765", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S4"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "bf1dfd9b854995ec8ba51cff051d35793afdffacd18b48cff91f7b39fb8a9ccd", "output_hash": "8d800485162553e438ca8ddff7de5613a4800ab0aeecee4f411e1cec8e2008bd"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "eb1602b81815b8e725c12cca", "attempt_id": "b8d91cb856c7", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:24:52.319585Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1677}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 1789, "message_tokens": 1701, "message_utf8_bytes": 6315, "cache_tensor_bytes": null, "serialized_payload_bytes": 6315, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 23379.230667022057, "ttft_ms": 13.726374949328601, "decode_ms": 122.53766704816371, "receiver_prefill_ms": 185.95604202710092}, "total_wall_ms": 23704.3878340628, "ttft_ms": 13.726374949328601, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7,730,705", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S4"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "bf1dfd9b854995ec8ba51cff051d35793afdffacd18b48cff91f7b39fb8a9ccd", "output_hash": "2e0f40d8945ba6021c3c207215c3d834da5944796960b093ea104a759c7b3bd0"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "51a9288e9694dc4897374c73", "attempt_id": "6401d41399ce", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:24:52.663643Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1677}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 1797, "message_tokens": 1701, "message_utf8_bytes": 6315, "cache_tensor_bytes": null, "serialized_payload_bytes": 6315, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 23379.230667022057, "ttft_ms": 13.569333008490503, "decode_ms": 54.109707940369844, "receiver_prefill_ms": 193.085000035353}, "total_wall_ms": 23642.942083999515, "ttft_ms": 13.569333008490503, "input_usage": {}, "output_usage": {"output_tokens": 4}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Melruk-3", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S4"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "bf1dfd9b854995ec8ba51cff051d35793afdffacd18b48cff91f7b39fb8a9ccd", "output_hash": "29e02c51d1e6eeabd8d3fcec3b482c44c02ebc350a5d4b73304fcba97294db93"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "1c57b1ab04800f5fe2a94e02", "attempt_id": "daed4660ed00", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:24:52.944602Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1677}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 1791, "message_tokens": 1701, "message_utf8_bytes": 6315, "cache_tensor_bytes": null, "serialized_payload_bytes": 6315, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 23379.230667022057, "ttft_ms": 13.64741602446884, "decode_ms": 109.05704193282872, "receiver_prefill_ms": 187.1469171019271}, "total_wall_ms": 23691.961333039217, "ttft_ms": 13.64741602446884, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "20481494", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S4"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "bf1dfd9b854995ec8ba51cff051d35793afdffacd18b48cff91f7b39fb8a9ccd", "output_hash": "9d5abea7e2661ec79dfef72bd0d183cd72ed188689c10073a76f648abf169abf"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "0363a8b8d4ddca2e4968857a", "attempt_id": "8f5854c7cdb9", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:24:53.274456Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1677}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 1792, "message_tokens": 1701, "message_utf8_bytes": 6315, "cache_tensor_bytes": null, "serialized_payload_bytes": 6315, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 23379.230667022057, "ttft_ms": 13.58570798765868, "decode_ms": 14.308791956864297, "receiver_prefill_ms": 193.83641704916954}, "total_wall_ms": 23603.843625984155, "ttft_ms": 13.58570798765868, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S4"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "bf1dfd9b854995ec8ba51cff051d35793afdffacd18b48cff91f7b39fb8a9ccd", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "8a2a1506d4709737926792db", "attempt_id": "c53da7cbad67", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:24:53.516392Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1677}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 1793, "message_tokens": 1701, "message_utf8_bytes": 6315, "cache_tensor_bytes": null, "serialized_payload_bytes": 6315, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 23379.230667022057, "ttft_ms": 14.23129194881767, "decode_ms": 13.558416976593435, "receiver_prefill_ms": 200.81454096361995}, "total_wall_ms": 23610.73416704312, "ttft_ms": 14.23129194881767, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S4"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "bf1dfd9b854995ec8ba51cff051d35793afdffacd18b48cff91f7b39fb8a9ccd", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "5b24ac2557ff43303f5cf646", "attempt_id": "85800351fc9d", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:25:20.285534Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 29}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 1202, "message_tokens": 1113, "message_utf8_bytes": 2987, "cache_tensor_bytes": null, "serialized_payload_bytes": 2987, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26520.410542027093, "ttft_ms": 13.765583047643304, "decode_ms": 121.388208004646, "receiver_prefill_ms": 152.4447079282254}, "total_wall_ms": 26810.068625956774, "ttft_ms": 13.765583047643304, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4,511,935", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S4"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "747cbfa9e09159527e79b5030dde7e23b2781f07b3aedfb601ba0642942f12fb", "output_hash": "3a6662e1122f9912388636d0e6506f4bda89215a4961a1ea0574efb588a05047"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "70d44ba4b873b6b1f9577ae0", "attempt_id": "3f38eb3cd5db", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:25:20.595313Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 29}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 1204, "message_tokens": 1113, "message_utf8_bytes": 2987, "cache_tensor_bytes": null, "serialized_payload_bytes": 2987, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26520.410542027093, "ttft_ms": 13.834709068760276, "decode_ms": 147.40666595753282, "receiver_prefill_ms": 792.1068749856204}, "total_wall_ms": 27475.66970891785, "ttft_ms": 13.834709068760276, "input_usage": {}, "output_usage": {"output_tokens": 11}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "JX-64S-4555", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S4"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "747cbfa9e09159527e79b5030dde7e23b2781f07b3aedfb601ba0642942f12fb", "output_hash": "f6f8ca3da45dc6c01a3dede99eab98f0f3fd2d2c5e105791d9c68c607877ed22"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "f869b2271be8ef84c61f6afb", "attempt_id": "44a4ce0fb169", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:25:21.568776Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 29}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 1210, "message_tokens": 1113, "message_utf8_bytes": 2987, "cache_tensor_bytes": null, "serialized_payload_bytes": 2987, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26520.410542027093, "ttft_ms": 14.015499968081713, "decode_ms": 97.33570797834545, "receiver_prefill_ms": 1040.7460830174387}, "total_wall_ms": 27674.47704204824, "ttft_ms": 14.015499968081713, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8859765", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S4"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "747cbfa9e09159527e79b5030dde7e23b2781f07b3aedfb601ba0642942f12fb", "output_hash": "8d800485162553e438ca8ddff7de5613a4800ab0aeecee4f411e1cec8e2008bd"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "f12ad7bd2689dcac693aafc3", "attempt_id": "6b53c2a77c8b", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:25:22.740383Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 29}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 1205, "message_tokens": 1113, "message_utf8_bytes": 2987, "cache_tensor_bytes": null, "serialized_payload_bytes": 2987, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26520.410542027093, "ttft_ms": 13.66258307825774, "decode_ms": 94.25008308608085, "receiver_prefill_ms": 127.36725003924221}, "total_wall_ms": 26757.666500052437, "ttft_ms": 13.66258307825774, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4135500", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S4"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "747cbfa9e09159527e79b5030dde7e23b2781f07b3aedfb601ba0642942f12fb", "output_hash": "6b5630c7cd8c2ad9a5337e78631ba45cc68e9c5436ea5150720f9dd9bfb6efeb"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "5740235f81a241df9985bd33", "attempt_id": "c8dd03611066", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:25:23.001509Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 29}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 1213, "message_tokens": 1113, "message_utf8_bytes": 2987, "cache_tensor_bytes": null, "serialized_payload_bytes": 2987, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26520.410542027093, "ttft_ms": 13.417250011116266, "decode_ms": 93.88745902106166, "receiver_prefill_ms": 126.45504192914814}, "total_wall_ms": 26756.021333043464, "ttft_ms": 13.417250011116266, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Project Ixbanruk-3", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S4"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "747cbfa9e09159527e79b5030dde7e23b2781f07b3aedfb601ba0642942f12fb", "output_hash": "4a52ee5b39aab55ee71eb97b7711e35dfa06dc12696c2a891362eb74c0d5207e"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "e4ab930c5788a997d0639ef4", "attempt_id": "f309d22631ab", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:25:23.254777Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 29}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 1207, "message_tokens": 1113, "message_utf8_bytes": 2987, "cache_tensor_bytes": null, "serialized_payload_bytes": 2987, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26520.410542027093, "ttft_ms": 13.695707893930376, "decode_ms": 110.66654196474701, "receiver_prefill_ms": 142.8877910366282}, "total_wall_ms": 26789.54279201571, "ttft_ms": 13.695707893930376, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "12241.22", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S4"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "747cbfa9e09159527e79b5030dde7e23b2781f07b3aedfb601ba0642942f12fb", "output_hash": "006931ee535eb411cabd77f6d126839bfffa0d95467d27b4136651e9c49e8dcc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "9f30d57aea26246b9b8d4454", "attempt_id": "7a67133f5397", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:25:23.541982Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 29}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 1208, "message_tokens": 1113, "message_utf8_bytes": 2987, "cache_tensor_bytes": null, "serialized_payload_bytes": 2987, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26520.410542027093, "ttft_ms": 13.846542104147375, "decode_ms": 13.605790911242366, "receiver_prefill_ms": 127.07837508060038}, "total_wall_ms": 26676.88600101974, "ttft_ms": 13.846542104147375, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S4"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "747cbfa9e09159527e79b5030dde7e23b2781f07b3aedfb601ba0642942f12fb", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "f63898df23131daf9a0dbd96", "attempt_id": "7458378380c4", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:25:23.716424Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 29}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 1209, "message_tokens": 1113, "message_utf8_bytes": 2987, "cache_tensor_bytes": null, "serialized_payload_bytes": 2987, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26520.410542027093, "ttft_ms": 13.833209057338536, "decode_ms": 13.417958049103618, "receiver_prefill_ms": 126.33054098114371}, "total_wall_ms": 26675.948542077094, "ttft_ms": 13.833209057338536, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S4"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "747cbfa9e09159527e79b5030dde7e23b2781f07b3aedfb601ba0642942f12fb", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "767d99c0e9e8622968d908f3", "attempt_id": "7bf1b9244d0f", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:25:50.778134Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 2523, "message_tokens": 2435, "message_utf8_bytes": 6413, "cache_tensor_bytes": null, "serialized_payload_bytes": 6413, "side_information_bytes": 0, "fetched_bytes": 3402, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26888.076208066195, "ttft_ms": 13.58887495007366, "decode_ms": 95.46912496443838, "receiver_prefill_ms": 339.3963749986142}, "total_wall_ms": 27339.99450004194, "ttft_ms": 13.58887495007366, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8544550", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S4"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "d6ed45c4e8799ecab484b9b3255a5cf11e403a36f30281ae82d0f26a5ecdd4da", "output_hash": "7ea69da7ca13297bce5fac6a79b7f74d78548d728dc85a91c943d1c2c8172768"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "9c98bd0c116b48a5a9541a83", "attempt_id": "9cd1ff56b78a", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:25:51.251407Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 2431, "message_tokens": 2341, "message_utf8_bytes": 6262, "cache_tensor_bytes": null, "serialized_payload_bytes": 6262, "side_information_bytes": 0, "fetched_bytes": 3224, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26888.076208066195, "ttft_ms": 15.687250066548586, "decode_ms": 156.32220800034702, "receiver_prefill_ms": 268.3007079176605}, "total_wall_ms": 27332.409708062187, "ttft_ms": 15.687250066548586, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 11}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "JX-64S-4555", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S4"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "e3bf0ebd4720fa89725d12f075c2b7f52b7418060d7e130ffc3b3047a0791da9", "output_hash": "f6f8ca3da45dc6c01a3dede99eab98f0f3fd2d2c5e105791d9c68c607877ed22"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c0c592c6f077d16e659ce8c1", "attempt_id": "fc03a360ff99", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:25:51.713231Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 2544, "message_tokens": 2448, "message_utf8_bytes": 6438, "cache_tensor_bytes": null, "serialized_payload_bytes": 6438, "side_information_bytes": 0, "fetched_bytes": 3435, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26888.076208066195, "ttft_ms": 13.5238750372082, "decode_ms": 94.09354208037257, "receiver_prefill_ms": 272.66583300661296}, "total_wall_ms": 27271.796248969622, "ttft_ms": 13.5238750372082, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8859765", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S4"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "53730a438dd04b0dc24e95e40d099eafee059df7c6f19edf2d66bf557a1b2723", "output_hash": "8d800485162553e438ca8ddff7de5613a4800ab0aeecee4f411e1cec8e2008bd"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "1f87818bc9dd8fd7fa24041a", "attempt_id": "79b58e363455", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:25:52.114870Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 2554, "message_tokens": 2463, "message_utf8_bytes": 6432, "cache_tensor_bytes": null, "serialized_payload_bytes": 6432, "side_information_bytes": 0, "fetched_bytes": 3400, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26888.076208066195, "ttft_ms": 14.196708099916577, "decode_ms": 101.0192499961704, "receiver_prefill_ms": 321.00641692522913}, "total_wall_ms": 27328.256541048177, "ttft_ms": 14.196708099916577, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7730705", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S4"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "d51aa1fe2a5742939ae0666dfcc473ea400df4de38851092ca6477ec3d971a92", "output_hash": "6c885473fb45b44c6a425cd3cd307b4771e35351a0e2c17186a92d2de9c1fd42"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "439606f95d6f10c3fe77d29a", "attempt_id": "1e4975bdbfd0", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:25:52.572439Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 2479, "message_tokens": 2380, "message_utf8_bytes": 6356, "cache_tensor_bytes": null, "serialized_payload_bytes": 6356, "side_information_bytes": 0, "fetched_bytes": 3341, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26888.076208066195, "ttft_ms": 13.779791072010994, "decode_ms": 68.90887499321252, "receiver_prefill_ms": 285.1857499917969}, "total_wall_ms": 27259.479833068326, "ttft_ms": 13.779791072010994, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Ixbanruk-3", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S4"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "e2b137540b050874bf12be3f3a3691cdf223954f88be95b33f91fd94248fbb91", "output_hash": "1cc74509d5485de66115e1f6b591a4a84f89826fbba3b26cfbf7cfcfdc216bf5"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "e70d3bf95c32d9fae3f208b0", "attempt_id": "e275a476af6e", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:25:52.962136Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 2553, "message_tokens": 2460, "message_utf8_bytes": 6458, "cache_tensor_bytes": null, "serialized_payload_bytes": 6458, "side_information_bytes": 0, "fetched_bytes": 3442, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26888.076208066195, "ttft_ms": 13.95712501835078, "decode_ms": 109.83708407729864, "receiver_prefill_ms": 297.77166701387614}, "total_wall_ms": 27313.128958106972, "ttft_ms": 13.95712501835078, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "19481404", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S4"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "739e83c574a741472bef8ca979562304ce4f38b503e223dee734c42177f34e0f", "output_hash": "7d6899209731cb544f6e2e5dd7090eaaa4d0b08762381b8d095485e142441fae"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "d36e50416d8323854385a52e", "attempt_id": "ff2eceb6a5f3", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:25:53.404417Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 2471, "message_tokens": 2377, "message_utf8_bytes": 6380, "cache_tensor_bytes": null, "serialized_payload_bytes": 6380, "side_information_bytes": 0, "fetched_bytes": 3333, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26888.076208066195, "ttft_ms": 13.673666981048882, "decode_ms": 13.49762489553541, "receiver_prefill_ms": 296.0884579224512}, "total_wall_ms": 27214.598666061647, "ttft_ms": 13.673666981048882, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S4"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "027c7a7b1e75fdca60c0c6fd7d00862d125cd0c46e0abceeb5dd683016e3b1f5", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "4498549f33396572986e9ff0", "attempt_id": "712d479eb390", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:25:53.748389Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 2530, "message_tokens": 2435, "message_utf8_bytes": 6413, "cache_tensor_bytes": null, "serialized_payload_bytes": 6413, "side_information_bytes": 0, "fetched_bytes": 3402, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26888.076208066195, "ttft_ms": 13.92537506762892, "decode_ms": 13.51895893458277, "receiver_prefill_ms": 334.06379201915115}, "total_wall_ms": 27253.03216604516, "ttft_ms": 13.92537506762892, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S4"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "d61f6e4ba559b6ad599776ba1dde125886b7cce11359e052c543341aa80f3413", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "ac465b052e171b7d8d3f7a8f", "attempt_id": "4be5fe683596", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:25:55.474944Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 234934272}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 42, "message_tokens": null, "message_utf8_bytes": 30528, "cache_tensor_bytes": 234934272, "serialized_payload_bytes": 234934272, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 1342.9936249740422, "ttft_ms": 91.66204195935279, "decode_ms": 107.14562493376434, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1556.7814169917256, "ttft_ms": 91.66204195935279, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8544550", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S4"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "7ea69da7ca13297bce5fac6a79b7f74d78548d728dc85a91c943d1c2c8172768"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "07aa59223e29e988f93aa7db", "attempt_id": "9e927a4653d0", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:25:55.707200Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 234934272}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 44, "message_tokens": null, "message_utf8_bytes": 30528, "cache_tensor_bytes": 234934272, "serialized_payload_bytes": 234934272, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 1342.9936249740422, "ttft_ms": 70.51808305550367, "decode_ms": 153.04691600613296, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1570.7210829714313, "ttft_ms": 70.51808305550367, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "XT-58M-1436", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S4"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "88a42ad8cf9f73f734996854479d027d426515d2bbd638fc89c418555a6c0484"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "5551ce60610c502d84dd17e2", "attempt_id": "092c4fc4e327", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:25:55.953745Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 234934272}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 50, "message_tokens": null, "message_utf8_bytes": 30528, "cache_tensor_bytes": 234934272, "serialized_payload_bytes": 234934272, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 1342.9936249740422, "ttft_ms": 142.6464159740135, "decode_ms": 109.63470791466534, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1599.8131249798462, "ttft_ms": 142.6464159740135, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4135500", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S4"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "6b5630c7cd8c2ad9a5337e78631ba45cc68e9c5436ea5150720f9dd9bfb6efeb"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "9e3cf9e4b91792eae7c44a4f", "attempt_id": "fbaee1ca560f", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:25:56.229371Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 234934272}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 45, "message_tokens": null, "message_utf8_bytes": 30528, "cache_tensor_bytes": 234934272, "serialized_payload_bytes": 234934272, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 1342.9936249740422, "ttft_ms": 64.21041709836572, "decode_ms": 109.11858303006738, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1520.5572079867125, "ttft_ms": 64.21041709836572, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7730705", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S4"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "6c885473fb45b44c6a425cd3cd307b4771e35351a0e2c17186a92d2de9c1fd42"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a9f19bec9b70a3208b67e289", "attempt_id": "2cb5e628ef41", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:25:56.425456Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 234934272}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 53, "message_tokens": null, "message_utf8_bytes": 30528, "cache_tensor_bytes": 234934272, "serialized_payload_bytes": 234934272, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 1342.9936249740422, "ttft_ms": 179.80724992230535, "decode_ms": 77.19920796807855, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1604.2816669214517, "ttft_ms": 179.80724992230535, "input_usage": {}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Ixbanruk-3", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S4"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "1cc74509d5485de66115e1f6b591a4a84f89826fbba3b26cfbf7cfcfdc216bf5"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "291a5ac6c2bdf46d0313c81f", "attempt_id": "4afd92133ca5", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:25:56.706749Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 234934272}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 30528, "cache_tensor_bytes": 234934272, "serialized_payload_bytes": 234934272, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 1342.9936249740422, "ttft_ms": 91.37199993710965, "decode_ms": 120.81383296754211, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1559.5775409601629, "ttft_ms": 91.37199993710965, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "14046129", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S4"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "a94f4176f29077a3639f8b4b208d4df1aa5ce852107b8f35a82b6d893bda6400"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "91578b1197d85a4963d393f5", "attempt_id": "84c0321fc7a6", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:25:56.940753Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 234934272}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 48, "message_tokens": null, "message_utf8_bytes": 30528, "cache_tensor_bytes": 234934272, "serialized_payload_bytes": 234934272, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 1342.9936249740422, "ttft_ms": 79.1197499493137, "decode_ms": 15.385249978862703, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1442.24495801609, "ttft_ms": 79.1197499493137, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S4"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "04809f3c85b7ad048d24e129", "attempt_id": "2c405b129886", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:25:57.057621Z", "task_family": "synthetic", "document_id": "syn_03019_8000", "document_hash": "f8991a6ada3dde65b1427947e7dcf1f720fb02070a2dc500ecfd2e9bc6d8bf60", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03019_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 234934272}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5790, "query_tokens": 49, "message_tokens": null, "message_utf8_bytes": 30528, "cache_tensor_bytes": 234934272, "serialized_payload_bytes": 234934272, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 1342.9936249740422, "ttft_ms": 61.951917014084756, "decode_ms": 15.303834108635783, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1424.911041976884, "ttft_ms": 61.951917014084756, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Ostfel-3", "bogus_amount": "8,995,865", "source": "S4"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03019_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "9876bb72d54e047ab9bfaf91", "attempt_id": "5bbfd502b67e", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:25:58.158623Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 2327, "message_tokens": 2241, "message_utf8_bytes": 8132, "cache_tensor_bytes": null, "serialized_payload_bytes": 8132, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.031125033274292946, "ttft_ms": 13.891041977331042, "decode_ms": 123.77791607286781, "receiver_prefill_ms": 269.98145796824247}, "total_wall_ms": 411.2374170217663, "ttft_ms": 13.891041977331042, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1,505,050", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "165eba437d60ab817b1f07a46115d2fdc9a94333061032c0fc3dda754aff54df", "output_hash": "ce50815ff9cdaab72b318f821f80a5ad96dc257c6df51048d6fd9c743b4157c6"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "47789b53c65f7a3c42d26010", "attempt_id": "1ed4118b2e4c", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:25:58.583426Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 2328, "message_tokens": 2241, "message_utf8_bytes": 8132, "cache_tensor_bytes": null, "serialized_payload_bytes": 8132, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.031125033274292946, "ttft_ms": 13.706292025744915, "decode_ms": 148.4004999510944, "receiver_prefill_ms": 258.01016693003476}, "total_wall_ms": 423.7230410799384, "ttft_ms": 13.706292025744915, "input_usage": {}, "output_usage": {"output_tokens": 11}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "BQ-59Z-6753", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "165eba437d60ab817b1f07a46115d2fdc9a94333061032c0fc3dda754aff54df", "output_hash": "db250f072c856fe7ddf4bbbd16e20c71b8bad1eef5e4480acd0dac30f8ffe9cd"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "63e50dc187329bd703a76296", "attempt_id": "00dfa5fecad3", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:25:59.020857Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 2335, "message_tokens": 2241, "message_utf8_bytes": 8132, "cache_tensor_bytes": null, "serialized_payload_bytes": 8132, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.031125033274292946, "ttft_ms": 13.389665982685983, "decode_ms": 122.2741660894826, "receiver_prefill_ms": 257.64349999371916}, "total_wall_ms": 396.9754580175504, "ttft_ms": 13.389665982685983, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4,655,775", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "165eba437d60ab817b1f07a46115d2fdc9a94333061032c0fc3dda754aff54df", "output_hash": "5f09e362aca0b92ad7da0d7a80d5d1ab3c49651619aa4dad4d1e51c4087fde12"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b6032e93710af68214019ee0", "attempt_id": "febb8c627a4d", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:25:59.431601Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 2326, "message_tokens": 2241, "message_utf8_bytes": 8132, "cache_tensor_bytes": null, "serialized_payload_bytes": 8132, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.031125033274292946, "ttft_ms": 13.73674999922514, "decode_ms": 94.26087501924485, "receiver_prefill_ms": 257.0220419438556}, "total_wall_ms": 368.54262510314584, "ttft_ms": 13.73674999922514, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "6974845", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "165eba437d60ab817b1f07a46115d2fdc9a94333061032c0fc3dda754aff54df", "output_hash": "28cc9ebee1b3f7e803628a9d5fa8730b20a446620cf59ca3afb5fbf6cd001b1d"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "fee5bfeaa9aefec38869ade5", "attempt_id": "11b38d1bd9e3", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:25:59.813929Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 2334, "message_tokens": 2241, "message_utf8_bytes": 8132, "cache_tensor_bytes": null, "serialized_payload_bytes": 8132, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.031125033274292946, "ttft_ms": 13.706458965316415, "decode_ms": 68.7168330186978, "receiver_prefill_ms": 258.1510409945622}, "total_wall_ms": 344.32387503329664, "ttft_ms": 13.706458965316415, "input_usage": {}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Kovar-9", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "165eba437d60ab817b1f07a46115d2fdc9a94333061032c0fc3dda754aff54df", "output_hash": "c969b5f7547945067f5b07e9f822d0043126326f979d950a8b1b045057fbb818"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "faac501ba3177a87aab4c8a8", "attempt_id": "7c1018dc8220", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:26:00.171575Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 2330, "message_tokens": 2241, "message_utf8_bytes": 8132, "cache_tensor_bytes": null, "serialized_payload_bytes": 8132, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.031125033274292946, "ttft_ms": 13.901834026910365, "decode_ms": 135.38845791481435, "receiver_prefill_ms": 258.445292012766}, "total_wall_ms": 411.3285830244422, "ttft_ms": 13.901834026910365, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "24,587,551", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "165eba437d60ab817b1f07a46115d2fdc9a94333061032c0fc3dda754aff54df", "output_hash": "2023eaa6b70ef24869eb7ff58a566e24b6fdb81ee00ac6a4fcb71cb6a57b2ce7"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "8a8436b02e73d301b9602024", "attempt_id": "5c384a3d7971", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:26:00.597057Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 2330, "message_tokens": 2241, "message_utf8_bytes": 8132, "cache_tensor_bytes": null, "serialized_payload_bytes": 8132, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.031125033274292946, "ttft_ms": 13.678541989065707, "decode_ms": 13.425040990114212, "receiver_prefill_ms": 257.3885830352083}, "total_wall_ms": 287.9697079770267, "ttft_ms": 13.678541989065707, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "165eba437d60ab817b1f07a46115d2fdc9a94333061032c0fc3dda754aff54df", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "88dd302b07d06f73fda012a1", "attempt_id": "df3bd4d8ab8a", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:26:00.899503Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 2332, "message_tokens": 2241, "message_utf8_bytes": 8132, "cache_tensor_bytes": null, "serialized_payload_bytes": 8132, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.031125033274292946, "ttft_ms": 13.64029198884964, "decode_ms": 13.615750009194016, "receiver_prefill_ms": 257.90862494613975}, "total_wall_ms": 288.8340000063181, "ttft_ms": 13.64029198884964, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "165eba437d60ab817b1f07a46115d2fdc9a94333061032c0fc3dda754aff54df", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "ddf596fa68a398f0b20147c9", "attempt_id": "fd3f8461fb19", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:26:15.278300Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 1125, "message_tokens": 1037, "message_utf8_bytes": 2464, "cache_tensor_bytes": null, "serialized_payload_bytes": 2464, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14076.079874997959, "ttft_ms": 13.453082996420562, "decode_ms": 121.43866694532335, "receiver_prefill_ms": 121.597999939695}, "total_wall_ms": 14334.436457953416, "ttft_ms": 13.453082996420562, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1,505,050", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "27038d9bff3c879c939b89b7ea5d25a4c137081173a39d93ac66531bf735bf88", "output_hash": "ce50815ff9cdaab72b318f821f80a5ad96dc257c6df51048d6fd9c743b4157c6"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "91dc128f2de9767a07332616", "attempt_id": "34e76b6361f4", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:26:15.551236Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 1126, "message_tokens": 1037, "message_utf8_bytes": 2464, "cache_tensor_bytes": null, "serialized_payload_bytes": 2464, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14076.079874997959, "ttft_ms": 13.444541953504086, "decode_ms": 148.56674999464303, "receiver_prefill_ms": 121.53533298987895}, "total_wall_ms": 14361.460041021928, "ttft_ms": 13.444541953504086, "input_usage": {}, "output_usage": {"output_tokens": 11}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "BQ-59Z-6753", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "27038d9bff3c879c939b89b7ea5d25a4c137081173a39d93ac66531bf735bf88", "output_hash": "db250f072c856fe7ddf4bbbd16e20c71b8bad1eef5e4480acd0dac30f8ffe9cd"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c1189560c1ddcbbf3762c9a8", "attempt_id": "41c1c71ca063", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:26:15.850679Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 1133, "message_tokens": 1037, "message_utf8_bytes": 2464, "cache_tensor_bytes": null, "serialized_payload_bytes": 2464, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14076.079874997959, "ttft_ms": 13.475790969096124, "decode_ms": 94.30174995213747, "receiver_prefill_ms": 121.63637496996671}, "total_wall_ms": 14307.325958041474, "ttft_ms": 13.475790969096124, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4655775", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "27038d9bff3c879c939b89b7ea5d25a4c137081173a39d93ac66531bf735bf88", "output_hash": "b7861433980fb1fe1626f95887719df5e18c1ad310548d7311301fbb11cb9171"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "1b9296a6414a748f3393b7d0", "attempt_id": "21fa5ee3c914", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:26:16.095831Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 1124, "message_tokens": 1037, "message_utf8_bytes": 2464, "cache_tensor_bytes": null, "serialized_payload_bytes": 2464, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14076.079874997959, "ttft_ms": 13.532208045944571, "decode_ms": 93.80483301356435, "receiver_prefill_ms": 120.72929204441607}, "total_wall_ms": 14305.946832988411, "ttft_ms": 13.532208045944571, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "6974845", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "27038d9bff3c879c939b89b7ea5d25a4c137081173a39d93ac66531bf735bf88", "output_hash": "28cc9ebee1b3f7e803628a9d5fa8730b20a446620cf59ca3afb5fbf6cd001b1d"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c2a5dada9268850e76feaa42", "attempt_id": "0523cf6ad121", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:26:16.339712Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 1132, "message_tokens": 1037, "message_utf8_bytes": 2464, "cache_tensor_bytes": null, "serialized_payload_bytes": 2464, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14076.079874997959, "ttft_ms": 13.550624949857593, "decode_ms": 55.206999997608364, "receiver_prefill_ms": 121.23470904771239}, "total_wall_ms": 14268.301084055565, "ttft_ms": 13.550624949857593, "input_usage": {}, "output_usage": {"output_tokens": 4}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Quital-3", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "27038d9bff3c879c939b89b7ea5d25a4c137081173a39d93ac66531bf735bf88", "output_hash": "afbf4910f8901622fd15b86baf53c5a654cecf1e73ba179354f34a3e6df66253"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "d99c266d00b763774b0ccbc1", "attempt_id": "127846a73c4b", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:26:16.546423Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 1128, "message_tokens": 1037, "message_utf8_bytes": 2464, "cache_tensor_bytes": null, "serialized_payload_bytes": 2464, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14076.079874997959, "ttft_ms": 13.945249957032502, "decode_ms": 134.72279103007168, "receiver_prefill_ms": 121.09899998176843}, "total_wall_ms": 14347.741791978478, "ttft_ms": 13.945249957032502, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "15,108,595", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "27038d9bff3c879c939b89b7ea5d25a4c137081173a39d93ac66531bf735bf88", "output_hash": "b62b0617a027f1f669aac2ed1075f7173d5ad558a25d4fd4495d7ccd886f4122"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "8c7d2dcbb15a0bcce0e6d96a", "attempt_id": "fe6924cfdb64", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:26:16.831983Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 1128, "message_tokens": 1037, "message_utf8_bytes": 2464, "cache_tensor_bytes": null, "serialized_payload_bytes": 2464, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14076.079874997959, "ttft_ms": 13.486167066730559, "decode_ms": 13.617207994684577, "receiver_prefill_ms": 121.70170806348324}, "total_wall_ms": 14226.7588339746, "ttft_ms": 13.486167066730559, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "27038d9bff3c879c939b89b7ea5d25a4c137081173a39d93ac66531bf735bf88", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "2f4b96312b7eb2676853fecc", "attempt_id": "7ba7cb9507d8", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:26:16.996827Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 1130, "message_tokens": 1037, "message_utf8_bytes": 2464, "cache_tensor_bytes": null, "serialized_payload_bytes": 2464, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14076.079874997959, "ttft_ms": 13.535708072595298, "decode_ms": 13.184332987293601, "receiver_prefill_ms": 121.57137505710125}, "total_wall_ms": 14226.264083990827, "ttft_ms": 13.535708072595298, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "27038d9bff3c879c939b89b7ea5d25a4c137081173a39d93ac66531bf735bf88", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "60d6432e3e42f1cb4ad470bd", "attempt_id": "1c1c97a19d01", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:26:28.464802Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 13}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 583, "message_tokens": 492, "message_utf8_bytes": 1319, "cache_tensor_bytes": null, "serialized_payload_bytes": 1319, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11303.13858401496, "ttft_ms": 13.166208984330297, "decode_ms": 13.420666917227209, "receiver_prefill_ms": 69.48766601271927}, "total_wall_ms": 11401.299125980586, "ttft_ms": 13.166208984330297, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "90cd220d5cde6ca36734c9346279ef63919ef90f4f02b03c24c3e105ae002674", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c11e8e46d605f1fc1784dac4", "attempt_id": "1e46636414bd", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:26:28.577425Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 13}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 584, "message_tokens": 492, "message_utf8_bytes": 1319, "cache_tensor_bytes": null, "serialized_payload_bytes": 1319, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11303.13858401496, "ttft_ms": 13.473708997480571, "decode_ms": 13.729624915868044, "receiver_prefill_ms": 69.00283298455179}, "total_wall_ms": 11400.619875988923, "ttft_ms": 13.473708997480571, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "90cd220d5cde6ca36734c9346279ef63919ef90f4f02b03c24c3e105ae002674", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "ced908c087475b77ef2ede90", "attempt_id": "8357dec00d8e", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:26:28.689004Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 13}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 591, "message_tokens": 492, "message_utf8_bytes": 1319, "cache_tensor_bytes": null, "serialized_payload_bytes": 1319, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11303.13858401496, "ttft_ms": 13.383334036916494, "decode_ms": 13.373541994951665, "receiver_prefill_ms": 69.56508290022612}, "total_wall_ms": 11400.67116706632, "ttft_ms": 13.383334036916494, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "90cd220d5cde6ca36734c9346279ef63919ef90f4f02b03c24c3e105ae002674", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "2546603395602a4c2173f5ba", "attempt_id": "54818d856a3c", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:26:28.800779Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 13}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 582, "message_tokens": 492, "message_utf8_bytes": 1319, "cache_tensor_bytes": null, "serialized_payload_bytes": 1319, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11303.13858401496, "ttft_ms": 13.223040965385735, "decode_ms": 92.68841589801013, "receiver_prefill_ms": 69.00087499525398}, "total_wall_ms": 11479.26983400248, "ttft_ms": 13.223040965385735, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5646720", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "90cd220d5cde6ca36734c9346279ef63919ef90f4f02b03c24c3e105ae002674", "output_hash": "e1d129e860627cf4623671ed186cb4ebbf5d63169d3d0dc14a62e1da2af61326"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "5c10d75f2518468176cd5a8e", "attempt_id": "770fc02d0a69", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:26:28.990956Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 13}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 590, "message_tokens": 492, "message_utf8_bytes": 1319, "cache_tensor_bytes": null, "serialized_payload_bytes": 1319, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11303.13858401496, "ttft_ms": 13.810208067297935, "decode_ms": 52.675665938295424, "receiver_prefill_ms": 69.00245801080018}, "total_wall_ms": 11439.839833998121, "ttft_ms": 13.810208067297935, "input_usage": {}, "output_usage": {"output_tokens": 4}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Quital-3", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "90cd220d5cde6ca36734c9346279ef63919ef90f4f02b03c24c3e105ae002674", "output_hash": "afbf4910f8901622fd15b86baf53c5a654cecf1e73ba179354f34a3e6df66253"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "8fbbb8f2a0432fcd9974bc45", "attempt_id": "e27295d8463f", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:26:29.141881Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 13}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 586, "message_tokens": 492, "message_utf8_bytes": 1319, "cache_tensor_bytes": null, "serialized_payload_bytes": 1319, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11303.13858401496, "ttft_ms": 13.841875013895333, "decode_ms": 106.68595798779279, "receiver_prefill_ms": 69.47837502229959}, "total_wall_ms": 11494.38387597911, "ttft_ms": 13.841875013895333, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "10509125", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "90cd220d5cde6ca36734c9346279ef63919ef90f4f02b03c24c3e105ae002674", "output_hash": "28f1d1171bdf58f57971a293e7fcba893f349d43f16479dc116111e3188a7144"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "6ea995929108ea3f5b88fbc2", "attempt_id": "e8125c6d3476", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:26:29.347381Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 13}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 586, "message_tokens": 492, "message_utf8_bytes": 1319, "cache_tensor_bytes": null, "serialized_payload_bytes": 1319, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11303.13858401496, "ttft_ms": 13.503707945346832, "decode_ms": 13.431500061415136, "receiver_prefill_ms": 69.36870806384832}, "total_wall_ms": 11400.663334061392, "ttft_ms": 13.503707945346832, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "90cd220d5cde6ca36734c9346279ef63919ef90f4f02b03c24c3e105ae002674", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "f2e214ed4ef0ce0e8045cc27", "attempt_id": "56f3a911850a", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:26:29.458789Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 13}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 588, "message_tokens": 492, "message_utf8_bytes": 1319, "cache_tensor_bytes": null, "serialized_payload_bytes": 1319, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11303.13858401496, "ttft_ms": 13.568124966695905, "decode_ms": 13.532124925404787, "receiver_prefill_ms": 69.30270907469094}, "total_wall_ms": 11400.780500960536, "ttft_ms": 13.568124966695905, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "90cd220d5cde6ca36734c9346279ef63919ef90f4f02b03c24c3e105ae002674", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "fc5184d340806ff0fc332c02", "attempt_id": "cc1cc0a77aa3", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:26:40.849816Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 1228, "message_tokens": 1138, "message_utf8_bytes": 2942, "cache_tensor_bytes": null, "serialized_payload_bytes": 2942, "side_information_bytes": 0, "fetched_bytes": 1605, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11279.119875049219, "ttft_ms": 13.484333991073072, "decode_ms": 13.46104103140533, "receiver_prefill_ms": 128.72649997007102}, "total_wall_ms": 11436.806124984287, "ttft_ms": 13.484333991073072, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "7114c15317d01d85402a4a21c9de71c8e585cae3394de5825f03783370ca7d49", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "17eb741a2c682597964375d5", "attempt_id": "bbba5f686299", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:26:41.022056Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 1194, "message_tokens": 1103, "message_utf8_bytes": 2887, "cache_tensor_bytes": null, "serialized_payload_bytes": 2887, "side_information_bytes": 0, "fetched_bytes": 1539, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11279.119875049219, "ttft_ms": 13.465374941006303, "decode_ms": 13.417166075669229, "receiver_prefill_ms": 127.0954170031473}, "total_wall_ms": 11435.040250071324, "ttft_ms": 13.465374941006303, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "5e5d699257a4a35908d96ed77d3043a3d9d5745f8c4fbb334caeae7ebc75ed49", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "5fc5166bc5bd9edd612ded9d", "attempt_id": "739be6023545", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:26:41.192190Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 1236, "message_tokens": 1138, "message_utf8_bytes": 2942, "cache_tensor_bytes": null, "serialized_payload_bytes": 2942, "side_information_bytes": 0, "fetched_bytes": 1605, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11279.119875049219, "ttft_ms": 13.50754196755588, "decode_ms": 95.48512496985495, "receiver_prefill_ms": 128.54058295488358}, "total_wall_ms": 11518.71974999085, "ttft_ms": 13.50754196755588, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4655775", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a72cdd7d338f2a4a7d122a14205b59afa2ea733892d028a5e5ad386727b36ac0", "output_hash": "b7861433980fb1fe1626f95887719df5e18c1ad310548d7311301fbb11cb9171"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "317a948d43bafa8a4edb3fc5", "attempt_id": "3bdc831292cc", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:26:41.446403Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 1227, "message_tokens": 1138, "message_utf8_bytes": 2942, "cache_tensor_bytes": null, "serialized_payload_bytes": 2942, "side_information_bytes": 0, "fetched_bytes": 1605, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11279.119875049219, "ttft_ms": 13.648167019709945, "decode_ms": 94.78899999521673, "receiver_prefill_ms": 128.2340829493478}, "total_wall_ms": 11517.962042125873, "ttft_ms": 13.648167019709945, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5646720", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ed231bef830df9e63d3c316d92d57b1c2ba31daf1fa13d3f76f6b51c9e239568", "output_hash": "e1d129e860627cf4623671ed186cb4ebbf5d63169d3d0dc14a62e1da2af61326"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "993eb1f6e0ee3b1a60872fe8", "attempt_id": "6a635ad51156", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:26:41.700850Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 1223, "message_tokens": 1126, "message_utf8_bytes": 2935, "cache_tensor_bytes": null, "serialized_payload_bytes": 2935, "side_information_bytes": 0, "fetched_bytes": 1597, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11279.119875049219, "ttft_ms": 13.721292023546994, "decode_ms": 66.73220894299448, "receiver_prefill_ms": 127.96512502245605}, "total_wall_ms": 11489.490084117278, "ttft_ms": 13.721292023546994, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Project Kovar-9", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "364ec93fc6647db193b94745597166a50ad17690ead3a316de6b9053f4f604d8", "output_hash": "ba9eb7d7e7e2ba2e01e022d7883eb6c4146ddc97ec6cf76ef4e65473620e0e77"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a8b2e5a20f4386fc77bd9919", "attempt_id": "8915492ba916", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:26:41.925778Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 1231, "message_tokens": 1138, "message_utf8_bytes": 2942, "cache_tensor_bytes": null, "serialized_payload_bytes": 2942, "side_information_bytes": 0, "fetched_bytes": 1605, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11279.119875049219, "ttft_ms": 13.43724993057549, "decode_ms": 107.23004199098796, "receiver_prefill_ms": 128.40429204516113}, "total_wall_ms": 11530.154917039908, "ttft_ms": 13.43724993057549, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "10316180", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "c4bcdb3b919a3185099769ae04a17e7902f26e851768dc04ecf302667b5ac434", "output_hash": "219c5eba6a4c4a04e8e34eb92bdadda23439f54f46f125860a51606ab0da82a6"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c040ea6f62dfd3f212f58b9c", "attempt_id": "5510da5c0b6f", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:26:42.191469Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 1213, "message_tokens": 1120, "message_utf8_bytes": 2940, "cache_tensor_bytes": null, "serialized_payload_bytes": 2940, "side_information_bytes": 0, "fetched_bytes": 1590, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11279.119875049219, "ttft_ms": 13.619291945360601, "decode_ms": 13.252875069156289, "receiver_prefill_ms": 127.40220804698765}, "total_wall_ms": 11435.405625030398, "ttft_ms": 13.619291945360601, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "fa008cabcd1ee371ede262f2acf1ca7a3771003f48caa6ea61f69de33221ab0b", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "f247a65b1bb4e3b52b811968", "attempt_id": "8afcc1b0ace9", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:26:42.362054Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 1233, "message_tokens": 1138, "message_utf8_bytes": 2942, "cache_tensor_bytes": null, "serialized_payload_bytes": 2942, "side_information_bytes": 0, "fetched_bytes": 1605, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 11279.119875049219, "ttft_ms": 13.65391700528562, "decode_ms": 13.530874974094331, "receiver_prefill_ms": 128.1990420538932}, "total_wall_ms": 11436.461750068702, "ttft_ms": 13.65391700528562, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "7114c15317d01d85402a4a21c9de71c8e585cae3394de5825f03783370ca7d49", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "8536abea44c1c5ad8ef418a0", "attempt_id": "75e3956884dc", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:26:42.801223Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 85487616}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 44, "message_tokens": null, "message_utf8_bytes": 8106, "cache_tensor_bytes": 85487616, "serialized_payload_bytes": 85487616, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 267.29420898482203, "ttft_ms": 30.090500018559396, "decode_ms": 125.79962494783103, "receiver_prefill_ms": 0.0}, "total_wall_ms": 427.5292919483036, "ttft_ms": 30.090500018559396, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1,505,050", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "ce50815ff9cdaab72b318f821f80a5ad96dc257c6df51048d6fd9c743b4157c6"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "331f85d9f0a1ecca8d2e9df5", "attempt_id": "4d5854bbd38e", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:26:42.975660Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 85487616}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 45, "message_tokens": null, "message_utf8_bytes": 8106, "cache_tensor_bytes": 85487616, "serialized_payload_bytes": 85487616, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 267.29420898482203, "ttft_ms": 27.15362492017448, "decode_ms": 138.73183296527714, "receiver_prefill_ms": 0.0}, "total_wall_ms": 437.05733388196677, "ttft_ms": 27.15362492017448, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "NJ-83E-8738", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "221cb023a964476f081fadfc6302fcebc6235ba5bc056d90a62ab7434c3df4f2"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c2ee9a00b3eb28003cc78183", "attempt_id": "f2eaae7074e5", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:26:43.160168Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 85487616}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 52, "message_tokens": null, "message_utf8_bytes": 8106, "cache_tensor_bytes": 85487616, "serialized_payload_bytes": 85487616, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 267.29420898482203, "ttft_ms": 28.124416014179587, "decode_ms": 97.26916695944965, "receiver_prefill_ms": 0.0}, "total_wall_ms": 396.831416990608, "ttft_ms": 28.124416014179587, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4655775", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b7861433980fb1fe1626f95887719df5e18c1ad310548d7311301fbb11cb9171"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "cd12c9793a8d97bef894f950", "attempt_id": "a0160d1e8bef", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:26:43.304330Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 85487616}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 43, "message_tokens": null, "message_utf8_bytes": 8106, "cache_tensor_bytes": 85487616, "serialized_payload_bytes": 85487616, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 267.29420898482203, "ttft_ms": 28.347707935608923, "decode_ms": 97.07549994345754, "receiver_prefill_ms": 0.0}, "total_wall_ms": 396.6988760512322, "ttft_ms": 28.347707935608923, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "6974845", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "28cc9ebee1b3f7e803628a9d5fa8730b20a446620cf59ca3afb5fbf6cd001b1d"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "f118dc4d16f74dbad2cbc93a", "attempt_id": "8ba1ceeb8692", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:26:43.448054Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 85487616}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 51, "message_tokens": null, "message_utf8_bytes": 8106, "cache_tensor_bytes": 85487616, "serialized_payload_bytes": 85487616, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 267.29420898482203, "ttft_ms": 27.80191705096513, "decode_ms": 55.52512500435114, "receiver_prefill_ms": 0.0}, "total_wall_ms": 354.4746679253876, "ttft_ms": 27.80191705096513, "input_usage": {}, "output_usage": {"output_tokens": 4}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Quital-3", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "afbf4910f8901622fd15b86baf53c5a654cecf1e73ba179354f34a3e6df66253"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a85a7e2d5269c8b9a2f6bfff", "attempt_id": "41e81f6d6fba", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:26:43.549794Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 85487616}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 8106, "cache_tensor_bytes": 85487616, "serialized_payload_bytes": 85487616, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 267.29420898482203, "ttft_ms": 27.131790993735194, "decode_ms": 125.32545800786465, "receiver_prefill_ms": 0.0}, "total_wall_ms": 423.7302510300651, "ttft_ms": 27.131790993735194, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "12589.061", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "79d41e7dd4899f3b26e4e55554b519e8be5fd62a13e57fd58b3992bf122f15d0"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "d4a9ddbef2189c74e03ab0fe", "attempt_id": "8e17feb09ae5", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:26:43.720262Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 85487616}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 8106, "cache_tensor_bytes": 85487616, "serialized_payload_bytes": 85487616, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 267.29420898482203, "ttft_ms": 27.261874987743795, "decode_ms": 14.126416062936187, "receiver_prefill_ms": 0.0}, "total_wall_ms": 312.3641669517383, "ttft_ms": 27.261874987743795, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "33c4f0a28e4ffc34c09c4946", "attempt_id": "37b1ecad1d8a", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:26:43.779450Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 85487616}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 49, "message_tokens": null, "message_utf8_bytes": 8106, "cache_tensor_bytes": 85487616, "serialized_payload_bytes": 85487616, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 267.29420898482203, "ttft_ms": 25.328415911644697, "decode_ms": 13.97900003939867, "receiver_prefill_ms": 0.0}, "total_wall_ms": 310.57095900177956, "ttft_ms": 25.328415911644697, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "24d8d45a2464ba6f4a8a8380", "attempt_id": "9f5f5deaf39d", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:26:43.837239Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 2441, "message_tokens": 2355, "message_utf8_bytes": 8636, "cache_tensor_bytes": null, "serialized_payload_bytes": 8636, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.027417088858783245, "ttft_ms": 13.507958967238665, "decode_ms": 121.46545900031924, "receiver_prefill_ms": 269.13987507577986}, "total_wall_ms": 407.9107081051916, "ttft_ms": 13.507958967238665, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1,505,050", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "6f7e54724fbe954c4046c548a8cdd636a01ba04e8c9b78d76cd806a2876e5f4d", "output_hash": "ce50815ff9cdaab72b318f821f80a5ad96dc257c6df51048d6fd9c743b4157c6"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "dc1b2b822efa65503921abf8", "attempt_id": "a7b769852cc9", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:26:44.259596Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 2442, "message_tokens": 2355, "message_utf8_bytes": 8636, "cache_tensor_bytes": null, "serialized_payload_bytes": 8636, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.027417088858783245, "ttft_ms": 14.074707985855639, "decode_ms": 149.0085000405088, "receiver_prefill_ms": 269.9879170395434}, "total_wall_ms": 436.8310001445934, "ttft_ms": 14.074707985855639, "input_usage": {}, "output_usage": {"output_tokens": 11}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "BQ-59Z-6753", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "6f7e54724fbe954c4046c548a8cdd636a01ba04e8c9b78d76cd806a2876e5f4d", "output_hash": "db250f072c856fe7ddf4bbbd16e20c71b8bad1eef5e4480acd0dac30f8ffe9cd"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "e1dfd735e71eb5329ba87313", "attempt_id": "d248d83bab3c", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:26:44.710756Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 2449, "message_tokens": 2355, "message_utf8_bytes": 8636, "cache_tensor_bytes": null, "serialized_payload_bytes": 8636, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.027417088858783245, "ttft_ms": 13.436875073239207, "decode_ms": 96.2450000224635, "receiver_prefill_ms": 268.2500840164721}, "total_wall_ms": 381.8036251468584, "ttft_ms": 13.436875073239207, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4655775", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "6f7e54724fbe954c4046c548a8cdd636a01ba04e8c9b78d76cd806a2876e5f4d", "output_hash": "b7861433980fb1fe1626f95887719df5e18c1ad310548d7311301fbb11cb9171"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "103a74ace7e434c06c387adb", "attempt_id": "8547f98e41d3", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:26:45.106839Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 2440, "message_tokens": 2355, "message_utf8_bytes": 8636, "cache_tensor_bytes": null, "serialized_payload_bytes": 8636, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.027417088858783245, "ttft_ms": 13.761833077296615, "decode_ms": 120.85820897482336, "receiver_prefill_ms": 268.30200001131743}, "total_wall_ms": 406.6082091303542, "ttft_ms": 13.761833077296615, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8,663,695", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "6f7e54724fbe954c4046c548a8cdd636a01ba04e8c9b78d76cd806a2876e5f4d", "output_hash": "8a8c707b1b5795e347ee48f254989477cc20a872192b710ac34c4cf564f0effa"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "25f137ca67a2e0afd6c45f62", "attempt_id": "56a5e9651b41", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:26:45.527797Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 2448, "message_tokens": 2355, "message_utf8_bytes": 8636, "cache_tensor_bytes": null, "serialized_payload_bytes": 8636, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.027417088858783245, "ttft_ms": 13.41895794030279, "decode_ms": 68.19254206493497, "receiver_prefill_ms": 267.7704159868881}, "total_wall_ms": 353.0967920087278, "ttft_ms": 13.41895794030279, "input_usage": {}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Kovar-9", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "6f7e54724fbe954c4046c548a8cdd636a01ba04e8c9b78d76cd806a2876e5f4d", "output_hash": "c969b5f7547945067f5b07e9f822d0043126326f979d950a8b1b045057fbb818"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "ab7ca2aceff12932acbb9272", "attempt_id": "51e0a05c863c", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:26:45.894682Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 2444, "message_tokens": 2355, "message_utf8_bytes": 8636, "cache_tensor_bytes": null, "serialized_payload_bytes": 8636, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.027417088858783245, "ttft_ms": 13.580249971710145, "decode_ms": 135.38854091893882, "receiver_prefill_ms": 268.083417089656}, "total_wall_ms": 420.71350011974573, "ttft_ms": 13.580249971710145, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "20,904,045", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "6f7e54724fbe954c4046c548a8cdd636a01ba04e8c9b78d76cd806a2876e5f4d", "output_hash": "32d189add228da1299c5d4903d642547467dbe76ab1fa46a08fe399c77b124a4"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "5f87e52442773dad365a0fa7", "attempt_id": "d21e1a8e39aa", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:26:46.329265Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 2444, "message_tokens": 2355, "message_utf8_bytes": 8636, "cache_tensor_bytes": null, "serialized_payload_bytes": 8636, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.027417088858783245, "ttft_ms": 13.551541953347623, "decode_ms": 13.78725003451109, "receiver_prefill_ms": 268.2530829915777}, "total_wall_ms": 299.3497920688242, "ttft_ms": 13.551541953347623, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "6f7e54724fbe954c4046c548a8cdd636a01ba04e8c9b78d76cd806a2876e5f4d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "4f728cc2776a251615188c8e", "attempt_id": "156e6db27b70", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:26:46.642770Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 2446, "message_tokens": 2355, "message_utf8_bytes": 8636, "cache_tensor_bytes": null, "serialized_payload_bytes": 8636, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.027417088858783245, "ttft_ms": 14.309375081211329, "decode_ms": 13.499709079042077, "receiver_prefill_ms": 274.38479103147984}, "total_wall_ms": 305.89258403051645, "ttft_ms": 14.309375081211329, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "6f7e54724fbe954c4046c548a8cdd636a01ba04e8c9b78d76cd806a2876e5f4d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "23cc30348a07c53a933d9a62", "attempt_id": "a5e2ff469b87", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:27:08.896800Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1609}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 1721, "message_tokens": 1634, "message_utf8_bytes": 5352, "cache_tensor_bytes": null, "serialized_payload_bytes": 5352, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 21932.983916951343, "ttft_ms": 13.465541996993124, "decode_ms": 13.449708931148052, "receiver_prefill_ms": 180.3730420069769}, "total_wall_ms": 22142.98708399292, "ttft_ms": 13.465541996993124, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a9cebf682785dc94dcdd0b202cacb245202933c222fc8019994325a1c059ab6f", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "e72aa891a124aa58ac4c0928", "attempt_id": "840af03861ca", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:27:09.121458Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1609}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 1722, "message_tokens": 1634, "message_utf8_bytes": 5352, "cache_tensor_bytes": null, "serialized_payload_bytes": 5352, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 21932.983916951343, "ttft_ms": 13.318500015884638, "decode_ms": 149.10033298656344, "receiver_prefill_ms": 180.86291593499482}, "total_wall_ms": 22278.976499917917, "ttft_ms": 13.318500015884638, "input_usage": {}, "output_usage": {"output_tokens": 11}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "BQ-59Z-6753", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a9cebf682785dc94dcdd0b202cacb245202933c222fc8019994325a1c059ab6f", "output_hash": "db250f072c856fe7ddf4bbbd16e20c71b8bad1eef5e4480acd0dac30f8ffe9cd"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "220b86f6643f21ce9d12c2cc", "attempt_id": "9ceb4fee8201", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:27:09.482028Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1609}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 1729, "message_tokens": 1634, "message_utf8_bytes": 5352, "cache_tensor_bytes": null, "serialized_payload_bytes": 5352, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 21932.983916951343, "ttft_ms": 13.54116597212851, "decode_ms": 94.4544579833746, "receiver_prefill_ms": 178.54995804373175}, "total_wall_ms": 22222.239958005957, "ttft_ms": 13.54116597212851, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4655775", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a9cebf682785dc94dcdd0b202cacb245202933c222fc8019994325a1c059ab6f", "output_hash": "b7861433980fb1fe1626f95887719df5e18c1ad310548d7311301fbb11cb9171"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "df8d23de959c6adea9819b4c", "attempt_id": "355f7e8f6990", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:27:09.785699Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1609}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 1720, "message_tokens": 1634, "message_utf8_bytes": 5352, "cache_tensor_bytes": null, "serialized_payload_bytes": 5352, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 21932.983916951343, "ttft_ms": 13.834208017215133, "decode_ms": 123.22387506719679, "receiver_prefill_ms": 179.7703750198707}, "total_wall_ms": 22252.690541907214, "ttft_ms": 13.834208017215133, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "6,974,845", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a9cebf682785dc94dcdd0b202cacb245202933c222fc8019994325a1c059ab6f", "output_hash": "79a7909b232c25342aa43beb14828205efddf9ba169e9a4732be6f886b0465ec"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "cefcd6bc17382eab746f39c2", "attempt_id": "7ed56389dfc9", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:27:10.119952Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1609}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 1728, "message_tokens": 1634, "message_utf8_bytes": 5352, "cache_tensor_bytes": null, "serialized_payload_bytes": 5352, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 21932.983916951343, "ttft_ms": 13.83949990849942, "decode_ms": 68.18979198578745, "receiver_prefill_ms": 179.76520804222673}, "total_wall_ms": 22197.576125035994, "ttft_ms": 13.83949990849942, "input_usage": {}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Kovar-9", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a9cebf682785dc94dcdd0b202cacb245202933c222fc8019994325a1c059ab6f", "output_hash": "c969b5f7547945067f5b07e9f822d0043126326f979d950a8b1b045057fbb818"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "ed4140b1abfb70324476ea06", "attempt_id": "bbb294cb4b57", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:27:10.398937Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1609}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 1724, "message_tokens": 1634, "message_utf8_bytes": 5352, "cache_tensor_bytes": null, "serialized_payload_bytes": 5352, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 21932.983916951343, "ttft_ms": 13.806959032081068, "decode_ms": 110.50708300899714, "receiver_prefill_ms": 180.76266604475677}, "total_wall_ms": 22240.936708985828, "ttft_ms": 13.806959032081068, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "11409075", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a9cebf682785dc94dcdd0b202cacb245202933c222fc8019994325a1c059ab6f", "output_hash": "5680edae8490840b49ab8e0f8f5363dd0d771f952818cb662d87507cc880b004"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "3c48ad5c526ab678b056c82b", "attempt_id": "e60d67f5dda4", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:27:10.721802Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1609}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 1724, "message_tokens": 1634, "message_utf8_bytes": 5352, "cache_tensor_bytes": null, "serialized_payload_bytes": 5352, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 21932.983916951343, "ttft_ms": 13.79187498241663, "decode_ms": 13.996249996125698, "receiver_prefill_ms": 188.08466591872275}, "total_wall_ms": 22151.78012487013, "ttft_ms": 13.79187498241663, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a9cebf682785dc94dcdd0b202cacb245202933c222fc8019994325a1c059ab6f", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "ac91751551322354f8fe9180", "attempt_id": "a8e0bf339723", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:27:10.955369Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1609}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 1726, "message_tokens": 1634, "message_utf8_bytes": 5352, "cache_tensor_bytes": null, "serialized_payload_bytes": 5352, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 21932.983916951343, "ttft_ms": 13.83866707328707, "decode_ms": 13.533624936826527, "receiver_prefill_ms": 198.73916695360094}, "total_wall_ms": 22161.949749919586, "ttft_ms": 13.83866707328707, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a9cebf682785dc94dcdd0b202cacb245202933c222fc8019994325a1c059ab6f", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "647778e2cbfc1939f6a66f7c", "attempt_id": "db31bf3d8512", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:27:39.250892Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 32}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 1323, "message_tokens": 1232, "message_utf8_bytes": 3205, "cache_tensor_bytes": null, "serialized_payload_bytes": 3205, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 28050.821875105612, "ttft_ms": 14.039249974302948, "decode_ms": 14.053374994546175, "receiver_prefill_ms": 848.9324999973178}, "total_wall_ms": 28930.883833090775, "ttft_ms": 14.039249974302948, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "0aab8a264fc2c9db61fe90748315b3a43cd278e4e07289df1501836a0af8d584", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "64a23da0eaa11c5491121f34", "attempt_id": "2bf3af816dbb", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:27:40.147582Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 32}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 1324, "message_tokens": 1232, "message_utf8_bytes": 3205, "cache_tensor_bytes": null, "serialized_payload_bytes": 3205, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 28050.821875105612, "ttft_ms": 14.019499998539686, "decode_ms": 154.2368340305984, "receiver_prefill_ms": 143.91250000335276}, "total_wall_ms": 28365.08995911572, "ttft_ms": 14.019499998539686, "input_usage": {}, "output_usage": {"output_tokens": 11}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "BQ-59Z-6753", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "0aab8a264fc2c9db61fe90748315b3a43cd278e4e07289df1501836a0af8d584", "output_hash": "db250f072c856fe7ddf4bbbd16e20c71b8bad1eef5e4480acd0dac30f8ffe9cd"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b5e5c739369b00dff284046a", "attempt_id": "5f47f739ca83", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:27:40.486546Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 32}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 1331, "message_tokens": 1232, "message_utf8_bytes": 3205, "cache_tensor_bytes": null, "serialized_payload_bytes": 3205, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 28050.821875105612, "ttft_ms": 13.506375020369887, "decode_ms": 122.74954200256616, "receiver_prefill_ms": 143.66587495896965}, "total_wall_ms": 28332.861917093396, "ttft_ms": 13.506375020369887, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4,655,775", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "0aab8a264fc2c9db61fe90748315b3a43cd278e4e07289df1501836a0af8d584", "output_hash": "5f09e362aca0b92ad7da0d7a80d5d1ab3c49651619aa4dad4d1e51c4087fde12"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "08aca3b52fca8e661b1bbcdc", "attempt_id": "5d1f439693b9", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:27:40.783237Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 32}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 1322, "message_tokens": 1232, "message_utf8_bytes": 3205, "cache_tensor_bytes": null, "serialized_payload_bytes": 3205, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 28050.821875105612, "ttft_ms": 13.521000044420362, "decode_ms": 121.39245797879994, "receiver_prefill_ms": 142.92891591321677}, "total_wall_ms": 28330.688083078712, "ttft_ms": 13.521000044420362, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4,655,775", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "0aab8a264fc2c9db61fe90748315b3a43cd278e4e07289df1501836a0af8d584", "output_hash": "5f09e362aca0b92ad7da0d7a80d5d1ab3c49651619aa4dad4d1e51c4087fde12"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "9e2a34cc7083c8ed1b0608bf", "attempt_id": "c5f4f6cfb3ed", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:27:41.077393Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 32}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 1330, "message_tokens": 1232, "message_utf8_bytes": 3205, "cache_tensor_bytes": null, "serialized_payload_bytes": 3205, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 28050.821875105612, "ttft_ms": 14.32404201477766, "decode_ms": 68.85029096156359, "receiver_prefill_ms": 144.644416985102}, "total_wall_ms": 28280.87541612331, "ttft_ms": 14.32404201477766, "input_usage": {}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Kovar-9", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "0aab8a264fc2c9db61fe90748315b3a43cd278e4e07289df1501836a0af8d584", "output_hash": "c969b5f7547945067f5b07e9f822d0043126326f979d950a8b1b045057fbb818"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "5df51f1c478dfa942d89340d", "attempt_id": "a87cdeb78183", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:27:41.322729Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 32}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 1326, "message_tokens": 1232, "message_utf8_bytes": 3205, "cache_tensor_bytes": null, "serialized_payload_bytes": 3205, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 28050.821875105612, "ttft_ms": 13.523042085580528, "decode_ms": 95.21950001362711, "receiver_prefill_ms": 142.82829104922712}, "total_wall_ms": 28304.46691601537, "ttft_ms": 13.523042085580528, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1898711", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "0aab8a264fc2c9db61fe90748315b3a43cd278e4e07289df1501836a0af8d584", "output_hash": "b4fe81063d0e205df4e89ead86b3c3d00d7c5d7d66c75749c9c70ec804cc00e8"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "e59fb37f832e16639d8fd8a3", "attempt_id": "83c38a2fa568", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:27:41.590833Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 32}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 1326, "message_tokens": 1232, "message_utf8_bytes": 3205, "cache_tensor_bytes": null, "serialized_payload_bytes": 3205, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 28050.821875105612, "ttft_ms": 13.563540996983647, "decode_ms": 13.63058399874717, "receiver_prefill_ms": 142.58133410476148}, "total_wall_ms": 28222.62454207521, "ttft_ms": 13.563540996983647, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "0aab8a264fc2c9db61fe90748315b3a43cd278e4e07289df1501836a0af8d584", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b522e7107c9f72ff5c34c5df", "attempt_id": "a2bdbc3a62c1", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:27:41.777425Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 32}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 1328, "message_tokens": 1232, "message_utf8_bytes": 3205, "cache_tensor_bytes": null, "serialized_payload_bytes": 3205, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 28050.821875105612, "ttft_ms": 13.62633309327066, "decode_ms": 13.705875026062131, "receiver_prefill_ms": 142.59266608860344}, "total_wall_ms": 28222.779625095427, "ttft_ms": 13.62633309327066, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "0aab8a264fc2c9db61fe90748315b3a43cd278e4e07289df1501836a0af8d584", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "9d41472c55f4d13f67218360", "attempt_id": "1c48852d231d", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:28:10.000009Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 2619, "message_tokens": 2529, "message_utf8_bytes": 6537, "cache_tensor_bytes": null, "serialized_payload_bytes": 6537, "side_information_bytes": 0, "fetched_bytes": 3333, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 28035.547083010897, "ttft_ms": 13.620582991279662, "decode_ms": 13.604499981738627, "receiver_prefill_ms": 342.5662500085309}, "total_wall_ms": 28408.829165971838, "ttft_ms": 13.620582991279662, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "5e7b46f13339bd6b22b905feb0fc34ca53c7feeb2f861c793eba2676d2981b3c", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "4494148375f88700fad8f180", "attempt_id": "65b03558fbf0", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:28:10.388605Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 2501, "message_tokens": 2410, "message_utf8_bytes": 6329, "cache_tensor_bytes": null, "serialized_payload_bytes": 6329, "side_information_bytes": 0, "fetched_bytes": 3108, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 28035.547083010897, "ttft_ms": 14.111083932220936, "decode_ms": 149.97862500604242, "receiver_prefill_ms": 284.52104108873755}, "total_wall_ms": 28487.955458112992, "ttft_ms": 14.111083932220936, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 11}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "BQ-59Z-6753", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "e484c6dd0347ce0aa4e929392e2633e442f1174fdcbd045668f5b229350af05c", "output_hash": "db250f072c856fe7ddf4bbbd16e20c71b8bad1eef5e4480acd0dac30f8ffe9cd"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "9e80234cef9da5653d0d275b", "attempt_id": "85ca8d7e1d68", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:28:10.891692Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 2616, "message_tokens": 2518, "message_utf8_bytes": 6510, "cache_tensor_bytes": null, "serialized_payload_bytes": 6510, "side_information_bytes": 0, "fetched_bytes": 3313, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 28035.547083010897, "ttft_ms": 13.492499943822622, "decode_ms": 95.65154102165252, "receiver_prefill_ms": 282.94274990912527}, "total_wall_ms": 28431.034333072603, "ttft_ms": 13.492499943822622, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4655775", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "2e60a5bac4e87c6c86dec000831eb5e3b771bc24f556057b501f4555cb2ecf21", "output_hash": "b7861433980fb1fe1626f95887719df5e18c1ad310548d7311301fbb11cb9171"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "1867369265aecee719263326", "attempt_id": "44b05fa1335c", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:28:11.302200Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 2618, "message_tokens": 2529, "message_utf8_bytes": 6537, "cache_tensor_bytes": null, "serialized_payload_bytes": 6537, "side_information_bytes": 0, "fetched_bytes": 3333, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 28035.547083010897, "ttft_ms": 13.889124966226518, "decode_ms": 94.91754195187241, "receiver_prefill_ms": 284.47145805694163}, "total_wall_ms": 28432.196208043024, "ttft_ms": 13.889124966226518, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5646720", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "7186018b6a3350f0f565b2c12083b151ad86c7829bc58576fa99624bf7e5bd85", "output_hash": "e1d129e860627cf4623671ed186cb4ebbf5d63169d3d0dc14a62e1da2af61326"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "0a0b324d166090ecdabee799", "attempt_id": "8df7a3fe9bf6", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:28:11.712942Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 2602, "message_tokens": 2505, "message_utf8_bytes": 6481, "cache_tensor_bytes": null, "serialized_payload_bytes": 6481, "side_information_bytes": 0, "fetched_bytes": 3262, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 28035.547083010897, "ttft_ms": 13.980792020447552, "decode_ms": 68.3760839747265, "receiver_prefill_ms": 317.11629196070135}, "total_wall_ms": 28438.77391703427, "ttft_ms": 13.980792020447552, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Kovar-9", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "3ca3cde20108ad319d8dc922123ea2fb5a2867194bce410b066b3bf7d6a0569b", "output_hash": "c969b5f7547945067f5b07e9f822d0043126326f979d950a8b1b045057fbb818"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "e8aedb0e47272a01b3a4f31e", "attempt_id": "296a2dfc0999", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:28:12.130782Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 2622, "message_tokens": 2529, "message_utf8_bytes": 6537, "cache_tensor_bytes": null, "serialized_payload_bytes": 6537, "side_information_bytes": 0, "fetched_bytes": 3333, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 28035.547083010897, "ttft_ms": 13.980207964777946, "decode_ms": 135.17216593027115, "receiver_prefill_ms": 335.47487505711615}, "total_wall_ms": 28523.67470797617, "ttft_ms": 13.980207964777946, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "19,448,440", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a77dad4d7a0729401a516c700a5134a9732c132577c137e12c850e2178335ab4", "output_hash": "cb22a425e0adaca6abcd3a046ec8e2446010ee27157675063eca3d020380d197"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "39ef6533a1152f20e689e42e", "attempt_id": "84749f5b440c", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:28:12.633691Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 2570, "message_tokens": 2477, "message_utf8_bytes": 6548, "cache_tensor_bytes": null, "serialized_payload_bytes": 6548, "side_information_bytes": 0, "fetched_bytes": 3304, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 28035.547083010897, "ttft_ms": 13.710000086575747, "decode_ms": 13.635792070999742, "receiver_prefill_ms": 428.7236660020426}, "total_wall_ms": 28495.703082997352, "ttft_ms": 13.710000086575747, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "3b118b05e0fb2e5caeb47aaa15f542169caf0e8a965b70694c277fbdcd6abbb6", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "55581f3062022a6ea68a4353", "attempt_id": "c0c1048b566a", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:28:13.108304Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 2629, "message_tokens": 2534, "message_utf8_bytes": 6548, "cache_tensor_bytes": null, "serialized_payload_bytes": 6548, "side_information_bytes": 0, "fetched_bytes": 3350, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 28035.547083010897, "ttft_ms": 13.793749967589974, "decode_ms": 13.334208051674068, "receiver_prefill_ms": 990.1274999137968}, "total_wall_ms": 29056.311708060093, "ttft_ms": 13.793749967589974, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a18eb01493865a50d0e2d8c2a121d303fab4c9ed44a213eae402d3537f1fd5ed", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "1f5b819e01bfe706227c2ba3", "attempt_id": "617283aff9e9", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:28:14.632407Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 92565504}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 44, "message_tokens": null, "message_utf8_bytes": 8582, "cache_tensor_bytes": 92565504, "serialized_payload_bytes": 92565504, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 487.7696660114452, "ttft_ms": 29.91020807530731, "decode_ms": 134.5039999578148, "receiver_prefill_ms": 0.0}, "total_wall_ms": 657.0747500518337, "ttft_ms": 29.91020807530731, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1,505,050", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "ce50815ff9cdaab72b318f821f80a5ad96dc257c6df51048d6fd9c743b4157c6"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "766f030d67e8377f2286e43d", "attempt_id": "78af82bbbcaf", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:28:14.818348Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 92565504}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 45, "message_tokens": null, "message_utf8_bytes": 8582, "cache_tensor_bytes": 92565504, "serialized_payload_bytes": 92565504, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 487.7696660114452, "ttft_ms": 27.635624981485307, "decode_ms": 142.99095899332315, "receiver_prefill_ms": 0.0}, "total_wall_ms": 661.1642490606755, "ttft_ms": 27.635624981485307, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "MH-96F-2099", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "95c230d14a055d30e0e1d4355caf07808e1c968b1feab9e0607d8bd37b1b43dc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "7b6bf03e022c02ac8f4a899d", "attempt_id": "445a22db73f3", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:28:15.007520Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 92565504}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 52, "message_tokens": null, "message_utf8_bytes": 8582, "cache_tensor_bytes": 92565504, "serialized_payload_bytes": 92565504, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 487.7696660114452, "ttft_ms": 29.698000056669116, "decode_ms": 98.41141698416322, "receiver_prefill_ms": 0.0}, "total_wall_ms": 618.524999008514, "ttft_ms": 29.698000056669116, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8955410", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "5b3f7267844b20dee460b7ecb0817bfd848295a2976bea23185c548209b23c6d"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b762f207ea2a590928e58774", "attempt_id": "3887c46c3a99", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:28:15.155750Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 92565504}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 43, "message_tokens": null, "message_utf8_bytes": 8582, "cache_tensor_bytes": 92565504, "serialized_payload_bytes": 92565504, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 487.7696660114452, "ttft_ms": 29.772916925139725, "decode_ms": 127.19450006261468, "receiver_prefill_ms": 0.0}, "total_wall_ms": 647.5418750196695, "ttft_ms": 29.772916925139725, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8,663,695", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "8a8c707b1b5795e347ee48f254989477cc20a872192b710ac34c4cf564f0effa"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "77ff29deca32088b5eec489f", "attempt_id": "d1ed57316219", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:28:15.330786Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 92565504}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 51, "message_tokens": null, "message_utf8_bytes": 8582, "cache_tensor_bytes": 92565504, "serialized_payload_bytes": 92565504, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 487.7696660114452, "ttft_ms": 29.239209019578993, "decode_ms": 70.78299997374415, "receiver_prefill_ms": 0.0}, "total_wall_ms": 590.0755409384146, "ttft_ms": 29.239209019578993, "input_usage": {}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Kovar-9", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "c969b5f7547945067f5b07e9f822d0043126326f979d950a8b1b045057fbb818"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "67449eaa3c6798cb2f566665", "attempt_id": "6b2347951ca1", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:28:15.448340Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 92565504}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 8582, "cache_tensor_bytes": 92565504, "serialized_payload_bytes": 92565504, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 487.7696660114452, "ttft_ms": 52.044957992620766, "decode_ms": 112.88833396974951, "receiver_prefill_ms": 0.0}, "total_wall_ms": 655.3992490516976, "ttft_ms": 52.044957992620766, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "15183105", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "2256f81e655ace749f49dfbb9d2b806fabd740ffda7236551018405cf9547092"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "6211333f512d03f041e3acdd", "attempt_id": "def3864623f7", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:28:15.630612Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 92565504}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 8582, "cache_tensor_bytes": 92565504, "serialized_payload_bytes": 92565504, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 487.7696660114452, "ttft_ms": 54.43337489850819, "decode_ms": 14.31937504094094, "receiver_prefill_ms": 0.0}, "total_wall_ms": 558.7156240362674, "ttft_ms": 54.43337489850819, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "66b4195ebabd4f87264ad831", "attempt_id": "5d64181577c4", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:28:15.715627Z", "task_family": "synthetic", "document_id": "syn_03021_2000", "document_hash": "3ac205b464552356699a4d2b0a7bf62497ead36e5bca73982e11ae5ee3810731", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03021_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 92565504}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1914, "query_tokens": 49, "message_tokens": null, "message_utf8_bytes": 8582, "cache_tensor_bytes": 92565504, "serialized_payload_bytes": 92565504, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 487.7696660114452, "ttft_ms": 26.682708063162863, "decode_ms": 14.19933291617781, "receiver_prefill_ms": 0.0}, "total_wall_ms": 531.5161660546437, "ttft_ms": 26.682708063162863, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Kovar-9", "bogus_amount": "5,251,445", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03021_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "e941fcd041296b2622a700d5", "attempt_id": "20e5816e8fe3", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:28:16.772212Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 6274, "message_tokens": 6185, "message_utf8_bytes": 30182, "cache_tensor_bytes": null, "serialized_payload_bytes": 30182, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.03845803439617157, "ttft_ms": 13.884666957892478, "decode_ms": 98.2951249461621, "receiver_prefill_ms": 752.3895830381662}, "total_wall_ms": 876.5390410553664, "ttft_ms": 13.884666957892478, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "687,975", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "30ea0ef6bbf3b28d57c376afa4b6b426001dcf7247cb72f5cc1167b02183b28d", "output_hash": "ef19034975881d08c60ac17454ae5ab6c3e5ce2b4b41f46e1509aa90f9a2ee93"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b35c1a8a657e227e6104d1ed", "attempt_id": "eb61f7760fac", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:28:17.666366Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 6271, "message_tokens": 6185, "message_utf8_bytes": 30182, "cache_tensor_bytes": null, "serialized_payload_bytes": 30182, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.03845803439617157, "ttft_ms": 13.802125002257526, "decode_ms": 139.92024993058294, "receiver_prefill_ms": 733.0049999291077}, "total_wall_ms": 896.4354160707444, "ttft_ms": 13.802125002257526, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "PF-86G-3220", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "30ea0ef6bbf3b28d57c376afa4b6b426001dcf7247cb72f5cc1167b02183b28d", "output_hash": "35ef227d6ed3e91b9e3ce606beff647cc9c529f914331f075b2ff73fb82a5fb2"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "877f3d2b4882f29da863c891", "attempt_id": "0333ce2e4650", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:28:18.580390Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 6278, "message_tokens": 6185, "message_utf8_bytes": 30182, "cache_tensor_bytes": null, "serialized_payload_bytes": 30182, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.03845803439617157, "ttft_ms": 13.902290957048535, "decode_ms": 124.60770807228982, "receiver_prefill_ms": 742.3407499445602}, "total_wall_ms": 890.6800409313291, "ttft_ms": 13.902290957048535, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "6,041,915", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "30ea0ef6bbf3b28d57c376afa4b6b426001dcf7247cb72f5cc1167b02183b28d", "output_hash": "f92c09b74703850def16f12d425c032d46efc5486dc44dfdda4d19d726710546"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "d7ba0fd8b70f8a4e357f7bf2", "attempt_id": "8b90cff29087", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:28:19.488718Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 6271, "message_tokens": 6185, "message_utf8_bytes": 30182, "cache_tensor_bytes": null, "serialized_payload_bytes": 30182, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.03845803439617157, "ttft_ms": 13.904000050388277, "decode_ms": 97.25762507878244, "receiver_prefill_ms": 736.6374999983236}, "total_wall_ms": 857.4109580367804, "ttft_ms": 13.904000050388277, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "314,635", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "30ea0ef6bbf3b28d57c376afa4b6b426001dcf7247cb72f5cc1167b02183b28d", "output_hash": "2d223f45af3560ed181ef7c63a9aeb2a1220ec6539ef00cd550bfbfb2dcef87b"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "866025e9305ad49e92b66fed", "attempt_id": "27ba2bf8af3a", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:28:20.362843Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 6281, "message_tokens": 6185, "message_utf8_bytes": 30182, "cache_tensor_bytes": null, "serialized_payload_bytes": 30182, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.03845803439617157, "ttft_ms": 14.08329198602587, "decode_ms": 83.49633391480893, "receiver_prefill_ms": 748.9049999276176}, "total_wall_ms": 857.6037919847295, "ttft_ms": 14.08329198602587, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Rifel-16", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "30ea0ef6bbf3b28d57c376afa4b6b426001dcf7247cb72f5cc1167b02183b28d", "output_hash": "34afeff5cc7f06df0544f826036c046cb8534ac4054132ee05dccf975102a81b"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "7339b499d0632370ed5e3b9e", "attempt_id": "4e95fbe9722d", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:28:21.238261Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 6274, "message_tokens": 6185, "message_utf8_bytes": 30182, "cache_tensor_bytes": null, "serialized_payload_bytes": 30182, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.03845803439617157, "ttft_ms": 13.89737508725375, "decode_ms": 111.61254195030779, "receiver_prefill_ms": 754.9543749773875}, "total_wall_ms": 890.384583035484, "ttft_ms": 13.89737508725375, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "22046.79", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "30ea0ef6bbf3b28d57c376afa4b6b426001dcf7247cb72f5cc1167b02183b28d", "output_hash": "a082dab7fa4264589e3930b23cc11d9c4d27ed51419fd26726bf7433c87d9a0c"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "61c746eb76f46fe9dc88cede", "attempt_id": "18bcf076898b", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:28:22.146178Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 6275, "message_tokens": 6185, "message_utf8_bytes": 30182, "cache_tensor_bytes": null, "serialized_payload_bytes": 30182, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.03845803439617157, "ttft_ms": 14.18875006493181, "decode_ms": 14.014750020578504, "receiver_prefill_ms": 765.0118339806795}, "total_wall_ms": 802.9102490982041, "ttft_ms": 14.18875006493181, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "30ea0ef6bbf3b28d57c376afa4b6b426001dcf7247cb72f5cc1167b02183b28d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b3b02d556dc9acc773968c1e", "attempt_id": "e877880565e1", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:28:22.966486Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 6276, "message_tokens": 6185, "message_utf8_bytes": 30182, "cache_tensor_bytes": null, "serialized_payload_bytes": 30182, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.03845803439617157, "ttft_ms": 14.081540983170271, "decode_ms": 14.046625001356006, "receiver_prefill_ms": 791.6234170552343}, "total_wall_ms": 829.4645419809967, "ttft_ms": 14.081540983170271, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "30ea0ef6bbf3b28d57c376afa4b6b426001dcf7247cb72f5cc1167b02183b28d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "f3025acd90c679c71ecd4cb7", "attempt_id": "c9342344a1c2", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:28:38.396330Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 1128, "message_tokens": 1037, "message_utf8_bytes": 2529, "cache_tensor_bytes": null, "serialized_payload_bytes": 2529, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14583.073709043674, "ttft_ms": 13.54591699782759, "decode_ms": 93.23150001000613, "receiver_prefill_ms": 121.63291696924716}, "total_wall_ms": 14813.404584070668, "ttft_ms": 13.54591699782759, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "687,975", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "3a4056e87072947fb029db964fab9955eba9ab8947bc294a1b2daecaa5c7490f", "output_hash": "ef19034975881d08c60ac17454ae5ab6c3e5ce2b4b41f46e1509aa90f9a2ee93"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "0ac754abdd04d7487f268b88", "attempt_id": "c3d3710ca4a4", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:28:38.644584Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 1125, "message_tokens": 1037, "message_utf8_bytes": 2529, "cache_tensor_bytes": null, "serialized_payload_bytes": 2529, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14583.073709043674, "ttft_ms": 13.755416963249445, "decode_ms": 134.27087501622736, "receiver_prefill_ms": 120.77479204162955}, "total_wall_ms": 14853.781500947662, "ttft_ms": 13.755416963249445, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "PF-86G-3220", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "3a4056e87072947fb029db964fab9955eba9ab8947bc294a1b2daecaa5c7490f", "output_hash": "35ef227d6ed3e91b9e3ce606beff647cc9c529f914331f075b2ff73fb82a5fb2"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "03f1d503403339c62919840a", "attempt_id": "bfd659c8e229", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:28:38.932307Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 1132, "message_tokens": 1037, "message_utf8_bytes": 2529, "cache_tensor_bytes": null, "serialized_payload_bytes": 2529, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14583.073709043674, "ttft_ms": 13.84849997702986, "decode_ms": 94.52579205390066, "receiver_prefill_ms": 121.38166697695851}, "total_wall_ms": 14814.717750996351, "ttft_ms": 13.84849997702986, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5774330", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "3a4056e87072947fb029db964fab9955eba9ab8947bc294a1b2daecaa5c7490f", "output_hash": "fb0fe829e6d55fb6fc250a6f0ea5faafc783e6dde33a8f87c31847eb8e1efd08"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "2c1ed6b33ed85b50d7e62278", "attempt_id": "04f71b66e1a9", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:28:39.181708Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 1125, "message_tokens": 1037, "message_utf8_bytes": 2529, "cache_tensor_bytes": null, "serialized_payload_bytes": 2529, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14583.073709043674, "ttft_ms": 13.48758302628994, "decode_ms": 94.17175000999123, "receiver_prefill_ms": 120.87945791427046}, "total_wall_ms": 14813.408751040697, "ttft_ms": 13.48758302628994, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "292,855", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "3a4056e87072947fb029db964fab9955eba9ab8947bc294a1b2daecaa5c7490f", "output_hash": "da1f2ed4aa6131f21ac7ec9afb1642c1e8f016fdb87da106a227d776c5349702"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "d565366f787b73ae86844375", "attempt_id": "a6dd96efa0c7", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:28:39.428936Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 1135, "message_tokens": 1037, "message_utf8_bytes": 2529, "cache_tensor_bytes": null, "serialized_payload_bytes": 2529, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14583.073709043674, "ttft_ms": 13.735540909692645, "decode_ms": 93.80545804742724, "receiver_prefill_ms": 121.6124159982428}, "total_wall_ms": 14814.098751056008, "ttft_ms": 13.735540909692645, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Talostmel-19", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "3a4056e87072947fb029db964fab9955eba9ab8947bc294a1b2daecaa5c7490f", "output_hash": "d36ce632b9874676c785373c6928c2f5d2c3ef3b6f327220429c713e0d2e5490"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "e6999ff21de74a99c6406b22", "attempt_id": "af0dc5da5870", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:28:39.676994Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 1128, "message_tokens": 1037, "message_utf8_bytes": 2529, "cache_tensor_bytes": null, "serialized_payload_bytes": 2529, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14583.073709043674, "ttft_ms": 13.482250040397048, "decode_ms": 94.74604192655534, "receiver_prefill_ms": 121.11800001002848}, "total_wall_ms": 14814.246251014993, "ttft_ms": 13.482250040397048, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "2304159", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "3a4056e87072947fb029db964fab9955eba9ab8947bc294a1b2daecaa5c7490f", "output_hash": "1f7bc9d1acc9966c962b11c4701f6a4f0012ad07abc9c24b016f5bcec862b573"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "557d10052e03add7c85fe935", "attempt_id": "3124173abebb", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:28:39.925835Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 1129, "message_tokens": 1037, "message_utf8_bytes": 2529, "cache_tensor_bytes": null, "serialized_payload_bytes": 2529, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14583.073709043674, "ttft_ms": 13.43970897141844, "decode_ms": 13.570999959483743, "receiver_prefill_ms": 121.40979198738933}, "total_wall_ms": 14733.315542107448, "ttft_ms": 13.43970897141844, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "3a4056e87072947fb029db964fab9955eba9ab8947bc294a1b2daecaa5c7490f", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "137f72b1d242d8a3349e08f9", "attempt_id": "9971a84e17c2", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:28:40.093585Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 1130, "message_tokens": 1037, "message_utf8_bytes": 2529, "cache_tensor_bytes": null, "serialized_payload_bytes": 2529, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14583.073709043674, "ttft_ms": 13.547584065236151, "decode_ms": 13.40558403171599, "receiver_prefill_ms": 121.14108295645565}, "total_wall_ms": 14733.019125065766, "ttft_ms": 13.547584065236151, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "3a4056e87072947fb029db964fab9955eba9ab8947bc294a1b2daecaa5c7490f", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "1d26244f8d5e2cd59d27e832", "attempt_id": "05df670815c8", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:28:54.936034Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 16}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 744, "message_tokens": 650, "message_utf8_bytes": 1769, "cache_tensor_bytes": null, "serialized_payload_bytes": 1769, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14674.899542005733, "ttft_ms": 13.682582997716963, "decode_ms": 95.3840829897672, "receiver_prefill_ms": 82.14020798914135}, "total_wall_ms": 14867.657250957564, "ttft_ms": 13.682582997716963, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "687,975", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "e7f7e10720e2759dc0c7e6cc62c24bf6663f89ed4f1bb9f6d1b6514400ed0bba", "output_hash": "ef19034975881d08c60ac17454ae5ab6c3e5ce2b4b41f46e1509aa90f9a2ee93"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "bec11631a188ea2b294709d0", "attempt_id": "5813535f9e5c", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:28:55.147823Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 16}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 741, "message_tokens": 650, "message_utf8_bytes": 1769, "cache_tensor_bytes": null, "serialized_payload_bytes": 1769, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14674.899542005733, "ttft_ms": 13.414792018011212, "decode_ms": 13.378959032706916, "receiver_prefill_ms": 80.96370904240757}, "total_wall_ms": 14784.114709007554, "ttft_ms": 13.414792018011212, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "e7f7e10720e2759dc0c7e6cc62c24bf6663f89ed4f1bb9f6d1b6514400ed0bba", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "e52c9ffb99fed920edf84d88", "attempt_id": "23bcc17c9dd4", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:28:55.275463Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 16}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 748, "message_tokens": 650, "message_utf8_bytes": 1769, "cache_tensor_bytes": null, "serialized_payload_bytes": 1769, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14674.899542005733, "ttft_ms": 13.505792012438178, "decode_ms": 95.94283404294401, "receiver_prefill_ms": 81.4348750282079}, "total_wall_ms": 14867.261082981713, "ttft_ms": 13.505792012438178, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5774330", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "e7f7e10720e2759dc0c7e6cc62c24bf6663f89ed4f1bb9f6d1b6514400ed0bba", "output_hash": "fb0fe829e6d55fb6fc250a6f0ea5faafc783e6dde33a8f87c31847eb8e1efd08"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "597b2ac704d450035cc973e5", "attempt_id": "97e2f49541ed", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:28:55.486010Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 16}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 741, "message_tokens": 650, "message_utf8_bytes": 1769, "cache_tensor_bytes": null, "serialized_payload_bytes": 1769, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14674.899542005733, "ttft_ms": 13.337415992282331, "decode_ms": 81.06137497816235, "receiver_prefill_ms": 80.57170896790922}, "total_wall_ms": 14851.268416969106, "ttft_ms": 13.337415992282331, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "292855", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "e7f7e10720e2759dc0c7e6cc62c24bf6663f89ed4f1bb9f6d1b6514400ed0bba", "output_hash": "4d2a03221fc71f99807c30668f5eb1a1311643cf5a57c7f7a36cc0c09df5f518"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "af3538fd21563fb5183db20c", "attempt_id": "067ea29031da", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:28:55.680305Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 16}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 751, "message_tokens": 650, "message_utf8_bytes": 1769, "cache_tensor_bytes": null, "serialized_payload_bytes": 1769, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14674.899542005733, "ttft_ms": 13.39183293748647, "decode_ms": 79.84487502835691, "receiver_prefill_ms": 81.98033296503127}, "total_wall_ms": 14851.535375928506, "ttft_ms": 13.39183293748647, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Rifel-16", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "e7f7e10720e2759dc0c7e6cc62c24bf6663f89ed4f1bb9f6d1b6514400ed0bba", "output_hash": "34afeff5cc7f06df0544f826036c046cb8534ac4054132ee05dccf975102a81b"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "38b23a43c098a28d7bfc2197", "attempt_id": "2c59310edeac", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:28:55.874471Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 16}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 744, "message_tokens": 650, "message_utf8_bytes": 1769, "cache_tensor_bytes": null, "serialized_payload_bytes": 1769, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14674.899542005733, "ttft_ms": 13.490624958649278, "decode_ms": 93.72395789250731, "receiver_prefill_ms": 80.55712503846735}, "total_wall_ms": 14864.132874994539, "ttft_ms": 13.490624958649278, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1034.92", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "e7f7e10720e2759dc0c7e6cc62c24bf6663f89ed4f1bb9f6d1b6514400ed0bba", "output_hash": "f7c59a0fb9a4fc03e78aede9c28a07034c4e923a03abea9d38dbe44b1600b492"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "e4ddee6421d8eb6ae6163708", "attempt_id": "4ef6ae768b3f", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:28:56.081088Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 16}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 745, "message_tokens": 650, "message_utf8_bytes": 1769, "cache_tensor_bytes": null, "serialized_payload_bytes": 1769, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14674.899542005733, "ttft_ms": 13.83487496059388, "decode_ms": 13.44624999910593, "receiver_prefill_ms": 80.75179101433605}, "total_wall_ms": 14784.353625029325, "ttft_ms": 13.83487496059388, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "e7f7e10720e2759dc0c7e6cc62c24bf6663f89ed4f1bb9f6d1b6514400ed0bba", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "732bff61662bd51632aa762b", "attempt_id": "70bd6363dd0f", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:28:56.208044Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 16}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 746, "message_tokens": 650, "message_utf8_bytes": 1769, "cache_tensor_bytes": null, "serialized_payload_bytes": 1769, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14674.899542005733, "ttft_ms": 13.562833075411618, "decode_ms": 13.21204099804163, "receiver_prefill_ms": 80.84770804271102}, "total_wall_ms": 14783.90645806212, "ttft_ms": 13.562833075411618, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "e7f7e10720e2759dc0c7e6cc62c24bf6663f89ed4f1bb9f6d1b6514400ed0bba", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "e1f6a4285ace70c13d9ce1b6", "attempt_id": "80a1d601e0aa", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:29:10.928705Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 1085, "message_tokens": 992, "message_utf8_bytes": 2621, "cache_tensor_bytes": null, "serialized_payload_bytes": 2621, "side_information_bytes": 0, "fetched_bytes": 1361, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14594.316959031858, "ttft_ms": 13.55112495366484, "decode_ms": 94.21429201029241, "receiver_prefill_ms": 116.37962504755706}, "total_wall_ms": 14823.182959109545, "ttft_ms": 13.55112495366484, "input_usage": {"lookups": 5}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "687,975", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "4b2910d797d15c5cb9074af211265a6aa24eec64996947e073a26d213fb8bcff", "output_hash": "ef19034975881d08c60ac17454ae5ab6c3e5ce2b4b41f46e1509aa90f9a2ee93"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a73c96153b7068fa21d28ebd", "attempt_id": "722a183aaff8", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:29:11.175102Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 1017, "message_tokens": 927, "message_utf8_bytes": 2528, "cache_tensor_bytes": null, "serialized_payload_bytes": 2528, "side_information_bytes": 0, "fetched_bytes": 1277, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14594.316959031858, "ttft_ms": 13.692917069420218, "decode_ms": 14.550624997355044, "receiver_prefill_ms": 107.08954103756696}, "total_wall_ms": 14731.966709019616, "ttft_ms": 13.692917069420218, "input_usage": {"lookups": 5}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ac43bc6034a7c07360f7b90add6726f2ac75f169e6a18f06659a39cd505d5d91", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "142df14e5824d53e5b485a4c", "attempt_id": "ac31c2088e57", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:29:11.332315Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 1089, "message_tokens": 992, "message_utf8_bytes": 2621, "cache_tensor_bytes": null, "serialized_payload_bytes": 2621, "side_information_bytes": 0, "fetched_bytes": 1361, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14594.316959031858, "ttft_ms": 13.411625055596232, "decode_ms": 13.53312493301928, "receiver_prefill_ms": 115.49337499309331}, "total_wall_ms": 14738.90233412385, "ttft_ms": 13.411625055596232, "input_usage": {"lookups": 5}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "8ac47f0cbefb82ca4ecd48085bcdfabfb6a42726b0319de6deaccd7675564603", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "5ffaf0650a298f04e1ac434c", "attempt_id": "4759e4bc20c5", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:29:11.493964Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 1082, "message_tokens": 992, "message_utf8_bytes": 2621, "cache_tensor_bytes": null, "serialized_payload_bytes": 2621, "side_information_bytes": 0, "fetched_bytes": 1361, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14594.316959031858, "ttft_ms": 13.45858396962285, "decode_ms": 80.31445904634893, "receiver_prefill_ms": 115.61437498312443}, "total_wall_ms": 14805.583918001503, "ttft_ms": 13.45858396962285, "input_usage": {"lookups": 5}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "753970", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "f63c7f5e9930bf7b4c2078a62a6d6e1c4fd9734d0c3e8d803ccb9efb80334dea", "output_hash": "ca9cbd16413b57561a8c177569d5245fb484b5a1ff07417932a75eab78f60c95"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "052ebf0e2fcedcda1508ce2a", "attempt_id": "6d4a595b6f4e", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:29:11.723310Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 1046, "message_tokens": 946, "message_utf8_bytes": 2543, "cache_tensor_bytes": null, "serialized_payload_bytes": 2543, "side_information_bytes": 0, "fetched_bytes": 1289, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14594.316959031858, "ttft_ms": 13.581290957517922, "decode_ms": 13.909499975852668, "receiver_prefill_ms": 114.03841699939221}, "total_wall_ms": 14737.622876069508, "ttft_ms": 13.581290957517922, "input_usage": {"lookups": 5}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "bc8bcf59cafd5c4f8caf42d0ae0086935dc13646780d50cbd092460af580850c", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "9ec8b51757e36333c3c87d72", "attempt_id": "e8e82a0643e0", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:29:11.884413Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 1085, "message_tokens": 992, "message_utf8_bytes": 2621, "cache_tensor_bytes": null, "serialized_payload_bytes": 2621, "side_information_bytes": 0, "fetched_bytes": 1361, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14594.316959031858, "ttft_ms": 13.44174996484071, "decode_ms": 95.00545798800886, "receiver_prefill_ms": 115.81654194742441}, "total_wall_ms": 14820.378041942604, "ttft_ms": 13.44174996484071, "input_usage": {"lookups": 5}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1441945", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "8ac47f0cbefb82ca4ecd48085bcdfabfb6a42726b0319de6deaccd7675564603", "output_hash": "ae63a4c42debd069f12befbd7586c0ce040e98a9efd95c9e1d0cb8a3c3a1114b"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "0e5d24d627587f7922746906", "attempt_id": "d7288e26cdc6", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:29:12.127479Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 1063, "message_tokens": 969, "message_utf8_bytes": 2613, "cache_tensor_bytes": null, "serialized_payload_bytes": 2613, "side_information_bytes": 0, "fetched_bytes": 1347, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14594.316959031858, "ttft_ms": 13.778125052340329, "decode_ms": 13.496334082446992, "receiver_prefill_ms": 115.65183300990611}, "total_wall_ms": 14739.11829199642, "ttft_ms": 13.778125052340329, "input_usage": {"lookups": 5}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "8086d90b17e79389f6c3aee84365cf30beafd27e7917c9966b0e61ae60ea8487", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "d4ccfa6f3fd836e84148e194", "attempt_id": "5c9038959e87", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:29:12.289723Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 1087, "message_tokens": 992, "message_utf8_bytes": 2621, "cache_tensor_bytes": null, "serialized_payload_bytes": 2621, "side_information_bytes": 0, "fetched_bytes": 1361, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14594.316959031858, "ttft_ms": 13.695167028345168, "decode_ms": 13.589290902018547, "receiver_prefill_ms": 115.83525000605732}, "total_wall_ms": 14739.213249995373, "ttft_ms": 13.695167028345168, "input_usage": {"lookups": 5}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "4b2910d797d15c5cb9074af211265a6aa24eec64996947e073a26d213fb8bcff", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "e5ea7ab2924114b6c035eb1c", "attempt_id": "ecce7c897099", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:29:13.228388Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 230879232}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 30156, "cache_tensor_bytes": 230879232, "serialized_payload_bytes": 230879232, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 776.5466249547899, "ttft_ms": 52.97391698695719, "decode_ms": 136.65237498935312, "receiver_prefill_ms": 0.0}, "total_wall_ms": 975.640874938108, "ttft_ms": 52.97391698695719, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7,120,879", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "c013fa074e4007a6fc0636d19a80febc8aee860429794ef4bbbe381185afe02c"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "f260f97179d0453e81858633", "attempt_id": "484ed88fc951", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:29:13.445437Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 230879232}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 44, "message_tokens": null, "message_utf8_bytes": 30156, "cache_tensor_bytes": 230879232, "serialized_payload_bytes": 230879232, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 776.5466249547899, "ttft_ms": 47.65170800965279, "decode_ms": 150.37654095795006, "receiver_prefill_ms": 0.0}, "total_wall_ms": 981.886291061528, "ttft_ms": 47.65170800965279, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "PF-86G-3220", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "35ef227d6ed3e91b9e3ce606beff647cc9c529f914331f075b2ff73fb82a5fb2"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "3a9aa45d26bd4001bf84ee87", "attempt_id": "4b97bacd614f", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:29:13.668323Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 230879232}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 51, "message_tokens": null, "message_utf8_bytes": 30156, "cache_tensor_bytes": 230879232, "serialized_payload_bytes": 230879232, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 776.5466249547899, "ttft_ms": 47.77066700626165, "decode_ms": 136.16691704373807, "receiver_prefill_ms": 0.0}, "total_wall_ms": 967.804417014122, "ttft_ms": 47.77066700626165, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "6,041,915", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "f92c09b74703850def16f12d425c032d46efc5486dc44dfdda4d19d726710546"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "3a4178682d4c0591a78135a3", "attempt_id": "d84f94d8877c", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:29:13.877549Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 230879232}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 44, "message_tokens": null, "message_utf8_bytes": 30156, "cache_tensor_bytes": 230879232, "serialized_payload_bytes": 230879232, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 776.5466249547899, "ttft_ms": 47.336916089989245, "decode_ms": 104.64612499345094, "receiver_prefill_ms": 0.0}, "total_wall_ms": 936.2797919893637, "ttft_ms": 47.336916089989245, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "314,635", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "2d223f45af3560ed181ef7c63a9aeb2a1220ec6539ef00cd550bfbfb2dcef87b"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b36db80b80269f06fbef1381", "attempt_id": "4423e407d43e", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:29:14.054840Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 230879232}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 54, "message_tokens": null, "message_utf8_bytes": 30156, "cache_tensor_bytes": 230879232, "serialized_payload_bytes": 230879232, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 776.5466249547899, "ttft_ms": 49.50524994637817, "decode_ms": 104.88683392759413, "receiver_prefill_ms": 0.0}, "total_wall_ms": 938.16637492273, "ttft_ms": 49.50524994637817, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Talostmel-19", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "d36ce632b9874676c785373c6928c2f5d2c3ef3b6f327220429c713e0d2e5490"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "70d7a290e30c60ddef3d186d", "attempt_id": "1874f852a560", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:29:14.233935Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 230879232}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 30156, "cache_tensor_bytes": 230879232, "serialized_payload_bytes": 230879232, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 776.5466249547899, "ttft_ms": 48.8192499615252, "decode_ms": 151.69762494042516, "receiver_prefill_ms": 0.0}, "total_wall_ms": 984.5911659067497, "ttft_ms": 48.8192499615252, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "20,821,812", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "78d202eed9ae00b429c5ee7f46f81cdc300290373c73658102933f966c5749e1"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "55adf1e07ce6f4afeb6c61ad", "attempt_id": "8d151e63d909", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:29:14.459865Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 230879232}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 48, "message_tokens": null, "message_utf8_bytes": 30156, "cache_tensor_bytes": 230879232, "serialized_payload_bytes": 230879232, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 776.5466249547899, "ttft_ms": 48.52995800320059, "decode_ms": 15.013708034530282, "receiver_prefill_ms": 0.0}, "total_wall_ms": 846.7708749230951, "ttft_ms": 48.52995800320059, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "385f8698e24e690c44f8cbca", "attempt_id": "e9acaaee3e7b", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:29:14.548215Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 230879232}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 49, "message_tokens": null, "message_utf8_bytes": 30156, "cache_tensor_bytes": 230879232, "serialized_payload_bytes": 230879232, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 776.5466249547899, "ttft_ms": 46.36929195839912, "decode_ms": 15.050416928716004, "receiver_prefill_ms": 0.0}, "total_wall_ms": 844.9183749035001, "ttft_ms": 46.36929195839912, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "3ae35f103b5336061e03cc58", "attempt_id": "efbdc0dd7c3a", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:29:14.634302Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 6388, "message_tokens": 6299, "message_utf8_bytes": 30686, "cache_tensor_bytes": null, "serialized_payload_bytes": 30686, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.029125018045306206, "ttft_ms": 13.961541000753641, "decode_ms": 96.22474992647767, "receiver_prefill_ms": 746.5892090694979}, "total_wall_ms": 871.6175830923021, "ttft_ms": 13.961541000753641, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "687,975", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a1629a65c5832d2a8c8d06c90d27e3d0c8d3c660c82ac7da12130c875c91992d", "output_hash": "ef19034975881d08c60ac17454ae5ab6c3e5ce2b4b41f46e1509aa90f9a2ee93"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "85a6ccd3ceec54b51d06ffd6", "attempt_id": "f0d586372060", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:29:15.522982Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 6385, "message_tokens": 6299, "message_utf8_bytes": 30686, "cache_tensor_bytes": null, "serialized_payload_bytes": 30686, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.029125018045306206, "ttft_ms": 14.104458969086409, "decode_ms": 138.4010420879349, "receiver_prefill_ms": 761.9426250457764}, "total_wall_ms": 924.1573330946267, "ttft_ms": 14.104458969086409, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "PF-86G-3220", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a1629a65c5832d2a8c8d06c90d27e3d0c8d3c660c82ac7da12130c875c91992d", "output_hash": "35ef227d6ed3e91b9e3ce606beff647cc9c529f914331f075b2ff73fb82a5fb2"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "fc9924e70a5ab115421163f4", "attempt_id": "185fc794b0dc", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:29:16.464122Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 6392, "message_tokens": 6299, "message_utf8_bytes": 30686, "cache_tensor_bytes": null, "serialized_payload_bytes": 30686, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.029125018045306206, "ttft_ms": 14.11649992223829, "decode_ms": 96.91716695670038, "receiver_prefill_ms": 795.9863749565557}, "total_wall_ms": 916.7706670705229, "ttft_ms": 14.11649992223829, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5774330", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a1629a65c5832d2a8c8d06c90d27e3d0c8d3c660c82ac7da12130c875c91992d", "output_hash": "fb0fe829e6d55fb6fc250a6f0ea5faafc783e6dde33a8f87c31847eb8e1efd08"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "cbb2237f23ca5a9d2d2fb934", "attempt_id": "ce6abd44bb1d", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:29:17.400572Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 6385, "message_tokens": 6299, "message_utf8_bytes": 30686, "cache_tensor_bytes": null, "serialized_payload_bytes": 30686, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.029125018045306206, "ttft_ms": 14.008625061251223, "decode_ms": 97.54433296620846, "receiver_prefill_ms": 777.63866703026}, "total_wall_ms": 898.926957976073, "ttft_ms": 14.008625061251223, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "314,635", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a1629a65c5832d2a8c8d06c90d27e3d0c8d3c660c82ac7da12130c875c91992d", "output_hash": "2d223f45af3560ed181ef7c63a9aeb2a1220ec6539ef00cd550bfbfb2dcef87b"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "f56cb21ebd6eb507803093e7", "attempt_id": "9bdeb9dab3f4", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:29:18.317017Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 6395, "message_tokens": 6299, "message_utf8_bytes": 30686, "cache_tensor_bytes": null, "serialized_payload_bytes": 30686, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.029125018045306206, "ttft_ms": 13.992166961543262, "decode_ms": 86.47658408153802, "receiver_prefill_ms": 793.4588340576738}, "total_wall_ms": 903.6159589886665, "ttft_ms": 13.992166961543262, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Rifel-16", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a1629a65c5832d2a8c8d06c90d27e3d0c8d3c660c82ac7da12130c875c91992d", "output_hash": "34afeff5cc7f06df0544f826036c046cb8534ac4054132ee05dccf975102a81b"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "3a36a9282714560395ae9804", "attempt_id": "11f6ece7d9c8", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:29:19.237824Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 6388, "message_tokens": 6299, "message_utf8_bytes": 30686, "cache_tensor_bytes": null, "serialized_payload_bytes": 30686, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.029125018045306206, "ttft_ms": 14.196209027431905, "decode_ms": 97.25879202596843, "receiver_prefill_ms": 805.4529159562662}, "total_wall_ms": 926.6065420815721, "ttft_ms": 14.196209027431905, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "3228899", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a1629a65c5832d2a8c8d06c90d27e3d0c8d3c660c82ac7da12130c875c91992d", "output_hash": "cfd0ed46facc7c79fb18dcb4d8d599dd6b3f146f49a8d86abe459544ef3b9cdb"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "493f9872a9fd0e622f1824e9", "attempt_id": "7f4a1a5a5a7a", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:29:20.181571Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 6389, "message_tokens": 6299, "message_utf8_bytes": 30686, "cache_tensor_bytes": null, "serialized_payload_bytes": 30686, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.029125018045306206, "ttft_ms": 14.219791977666318, "decode_ms": 13.836249941959977, "receiver_prefill_ms": 804.2557090520859}, "total_wall_ms": 841.9538750313222, "ttft_ms": 14.219791977666318, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a1629a65c5832d2a8c8d06c90d27e3d0c8d3c660c82ac7da12130c875c91992d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "67790b175a013bd5919fd1cd", "attempt_id": "d030a0349ca8", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:29:21.040347Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 6390, "message_tokens": 6299, "message_utf8_bytes": 30686, "cache_tensor_bytes": null, "serialized_payload_bytes": 30686, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.029125018045306206, "ttft_ms": 14.124833047389984, "decode_ms": 13.952166074886918, "receiver_prefill_ms": 819.719250081107}, "total_wall_ms": 857.5393341016024, "ttft_ms": 14.124833047389984, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a1629a65c5832d2a8c8d06c90d27e3d0c8d3c660c82ac7da12130c875c91992d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "cb4c086f4314e214a9005b9e", "attempt_id": "6e851f63a52f", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:29:47.291850Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1827}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 1942, "message_tokens": 1852, "message_utf8_bytes": 6509, "cache_tensor_bytes": null, "serialized_payload_bytes": 6509, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25376.599375042133, "ttft_ms": 13.657125062309206, "decode_ms": 94.81058397796005, "receiver_prefill_ms": 204.9281670479104}, "total_wall_ms": 25694.177916971967, "ttft_ms": 13.657125062309206, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "687,975", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "fce2d59b037adafbbe980681b5eb1030f888f8ba5b729d9f591bb94c1bfeb0d2", "output_hash": "ef19034975881d08c60ac17454ae5ab6c3e5ce2b4b41f46e1509aa90f9a2ee93"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "565726825dc80735de3ee34b", "attempt_id": "b443e5785223", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:29:47.627550Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1827}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 1939, "message_tokens": 1852, "message_utf8_bytes": 6509, "cache_tensor_bytes": null, "serialized_payload_bytes": 6509, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25376.599375042133, "ttft_ms": 13.4482909925282, "decode_ms": 134.8007080378011, "receiver_prefill_ms": 204.18279198929667}, "total_wall_ms": 25732.01112507377, "ttft_ms": 13.4482909925282, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "PF-86G-3220", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "fce2d59b037adafbbe980681b5eb1030f888f8ba5b729d9f591bb94c1bfeb0d2", "output_hash": "35ef227d6ed3e91b9e3ce606beff647cc9c529f914331f075b2ff73fb82a5fb2"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "1f359e3d7c870dbb8f3c1c5e", "attempt_id": "dfd07ddc7ae8", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:29:48.000770Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1827}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 1946, "message_tokens": 1852, "message_utf8_bytes": 6509, "cache_tensor_bytes": null, "serialized_payload_bytes": 6509, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25376.599375042133, "ttft_ms": 13.440750073641539, "decode_ms": 94.66208308003843, "receiver_prefill_ms": 205.138957942836}, "total_wall_ms": 25692.98383302521, "ttft_ms": 13.440750073641539, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "6041915", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "fce2d59b037adafbbe980681b5eb1030f888f8ba5b729d9f591bb94c1bfeb0d2", "output_hash": "5f8f3414199be3bad75f36df45e3759125ec38c5a7eccea16448f93268c5fa92"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "9949f335e7818f4b95c285b3", "attempt_id": "8760cea08ee4", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:29:48.334052Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1827}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 1939, "message_tokens": 1852, "message_utf8_bytes": 6509, "cache_tensor_bytes": null, "serialized_payload_bytes": 6509, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25376.599375042133, "ttft_ms": 13.66650010459125, "decode_ms": 94.48895894456655, "receiver_prefill_ms": 203.71045800857246}, "total_wall_ms": 25691.459250054322, "ttft_ms": 13.66650010459125, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "314,635", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "fce2d59b037adafbbe980681b5eb1030f888f8ba5b729d9f591bb94c1bfeb0d2", "output_hash": "2d223f45af3560ed181ef7c63a9aeb2a1220ec6539ef00cd550bfbfb2dcef87b"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "0e89cb7990f74680918b53d3", "attempt_id": "ee8b7e90c347", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:29:48.666845Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1827}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 1949, "message_tokens": 1852, "message_utf8_bytes": 6509, "cache_tensor_bytes": null, "serialized_payload_bytes": 6509, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25376.599375042133, "ttft_ms": 13.55487504042685, "decode_ms": 94.94716697372496, "receiver_prefill_ms": 206.4797499915585}, "total_wall_ms": 25694.618417066522, "ttft_ms": 13.55487504042685, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Talostmel-19", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "fce2d59b037adafbbe980681b5eb1030f888f8ba5b729d9f591bb94c1bfeb0d2", "output_hash": "d36ce632b9874676c785373c6928c2f5d2c3ef3b6f327220429c713e0d2e5490"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a6514d826baf5b138f85e2bc", "attempt_id": "3c5fec696c01", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:29:49.002278Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1827}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 1942, "message_tokens": 1852, "message_utf8_bytes": 6509, "cache_tensor_bytes": null, "serialized_payload_bytes": 6509, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25376.599375042133, "ttft_ms": 13.879333040677011, "decode_ms": 136.29879092331976, "receiver_prefill_ms": 213.41824997216463}, "total_wall_ms": 25743.202042067423, "ttft_ms": 13.879333040677011, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "22,077,139", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "fce2d59b037adafbbe980681b5eb1030f888f8ba5b729d9f591bb94c1bfeb0d2", "output_hash": "f9ed93d4a2e1fdf1782bdde030148563c3c5321ce15a2e420860c8ab8a868183"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "877f84bdc134299a0579cb7b", "attempt_id": "ad7efdb40c11", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:29:49.386284Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1827}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 1943, "message_tokens": 1852, "message_utf8_bytes": 6509, "cache_tensor_bytes": null, "serialized_payload_bytes": 6509, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25376.599375042133, "ttft_ms": 13.712832937017083, "decode_ms": 13.8938749441877, "receiver_prefill_ms": 211.3716669846326}, "total_wall_ms": 25618.57658298686, "ttft_ms": 13.712832937017083, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "fce2d59b037adafbbe980681b5eb1030f888f8ba5b729d9f591bb94c1bfeb0d2", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "2e613d30993b2166efc4ec84", "attempt_id": "6fbfa74651ad", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:29:49.645822Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1827}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 1944, "message_tokens": 1852, "message_utf8_bytes": 6509, "cache_tensor_bytes": null, "serialized_payload_bytes": 6509, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 25376.599375042133, "ttft_ms": 13.752958038821816, "decode_ms": 13.984875055029988, "receiver_prefill_ms": 223.85970794130117}, "total_wall_ms": 25631.163166020997, "ttft_ms": 13.752958038821816, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "fce2d59b037adafbbe980681b5eb1030f888f8ba5b729d9f591bb94c1bfeb0d2", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "50a4445f39f9db1c3fbea8c1", "attempt_id": "e97125b7c71d", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:30:16.057227Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 27}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 1179, "message_tokens": 1085, "message_utf8_bytes": 2921, "cache_tensor_bytes": null, "serialized_payload_bytes": 2921, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26139.297457993962, "ttft_ms": 13.463124982081354, "decode_ms": 93.7956249108538, "receiver_prefill_ms": 125.10354199912399}, "total_wall_ms": 26378.505249973387, "ttft_ms": 13.463124982081354, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "326,975", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "6af415f67866ccc4a833b68596dca13df4f4e8c7106c77693cbe69832bfb6d35", "output_hash": "f578533e5179bf25f29e177b2f8eec198dca384c7f5584e473eb11f7427728bf"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "09a0b22c9b201f586789a6e9", "attempt_id": "c069d1e2169e", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:30:16.314814Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 27}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 1176, "message_tokens": 1085, "message_utf8_bytes": 2921, "cache_tensor_bytes": null, "serialized_payload_bytes": 2921, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26139.297457993962, "ttft_ms": 14.675707905553281, "decode_ms": 14.280749950557947, "receiver_prefill_ms": 379.5532500371337}, "total_wall_ms": 26549.725457909517, "ttft_ms": 14.675707905553281, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "6af415f67866ccc4a833b68596dca13df4f4e8c7106c77693cbe69832bfb6d35", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "aa4b3ccfee1edc84f6ea7acf", "attempt_id": "f2e46f808e74", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:30:16.744500Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 27}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 1183, "message_tokens": 1085, "message_utf8_bytes": 2921, "cache_tensor_bytes": null, "serialized_payload_bytes": 2921, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26139.297457993962, "ttft_ms": 89.05462501570582, "decode_ms": 14.41008411347866, "receiver_prefill_ms": 1895.2112080296502}, "total_wall_ms": 28140.168583020568, "ttft_ms": 89.05462501570582, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "6af415f67866ccc4a833b68596dca13df4f4e8c7106c77693cbe69832bfb6d35", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "5abe3eed0dc5e529497e5488", "attempt_id": "4edd8c3131de", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:30:18.765642Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 27}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 1176, "message_tokens": 1085, "message_utf8_bytes": 2921, "cache_tensor_bytes": null, "serialized_payload_bytes": 2921, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26139.297457993962, "ttft_ms": 14.081125031225383, "decode_ms": 97.65345801133662, "receiver_prefill_ms": 144.03470803517848}, "total_wall_ms": 26397.122082998976, "ttft_ms": 14.081125031225383, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "326,975", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "6af415f67866ccc4a833b68596dca13df4f4e8c7106c77693cbe69832bfb6d35", "output_hash": "f578533e5179bf25f29e177b2f8eec198dca384c7f5584e473eb11f7427728bf"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "f3a4b5f1f12ea42f3758f1a1", "attempt_id": "796d5eafc483", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:30:19.047883Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 27}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 1186, "message_tokens": 1085, "message_utf8_bytes": 2921, "cache_tensor_bytes": null, "serialized_payload_bytes": 2921, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26139.297457993962, "ttft_ms": 14.015124994330108, "decode_ms": 94.56670901272446, "receiver_prefill_ms": 146.98204200249165}, "total_wall_ms": 26396.862208028324, "ttft_ms": 14.015124994330108, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Talostmel-19", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "6af415f67866ccc4a833b68596dca13df4f4e8c7106c77693cbe69832bfb6d35", "output_hash": "d36ce632b9874676c785373c6928c2f5d2c3ef3b6f327220429c713e0d2e5490"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "121dd36fb62e63768ce7b072", "attempt_id": "f29ff618c8a9", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:30:19.324079Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 27}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 1179, "message_tokens": 1085, "message_utf8_bytes": 2921, "cache_tensor_bytes": null, "serialized_payload_bytes": 2921, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26139.297457993962, "ttft_ms": 13.477791915647686, "decode_ms": 94.51912506483495, "receiver_prefill_ms": 154.128959053196}, "total_wall_ms": 26403.27995806001, "ttft_ms": 13.477791915647686, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1389750", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "6af415f67866ccc4a833b68596dca13df4f4e8c7106c77693cbe69832bfb6d35", "output_hash": "e0aabc7ef11c09c192fa67d678c29e2d671d7ded07b68555a8f7ee882c681adf"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "2820d3d7c8dff1b3a0afff8a", "attempt_id": "036506adc868", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:30:19.605643Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 27}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 1180, "message_tokens": 1085, "message_utf8_bytes": 2921, "cache_tensor_bytes": null, "serialized_payload_bytes": 2921, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26139.297457993962, "ttft_ms": 13.403709046542645, "decode_ms": 134.20887489337474, "receiver_prefill_ms": 124.29491698276252}, "total_wall_ms": 26413.080791011453, "ttft_ms": 13.403709046542645, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "ZH-49V-5200", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "6af415f67866ccc4a833b68596dca13df4f4e8c7106c77693cbe69832bfb6d35", "output_hash": "86e838b69725cb37b0278f6792e5169986b076b907a6e4269e72578c9f3d2a0a"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "4de8085d3d1e7fbdf5a8a6e0", "attempt_id": "e5fb021f0268", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:30:19.896851Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 27}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 1181, "message_tokens": 1085, "message_utf8_bytes": 2921, "cache_tensor_bytes": null, "serialized_payload_bytes": 2921, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26139.297457993962, "ttft_ms": 13.628540909849107, "decode_ms": 13.364666025154293, "receiver_prefill_ms": 124.47062495630234}, "total_wall_ms": 26292.649833019823, "ttft_ms": 13.628540909849107, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "6af415f67866ccc4a833b68596dca13df4f4e8c7106c77693cbe69832bfb6d35", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "de4f59eff4ccafb2d4dad690", "attempt_id": "a1fa09571d67", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:30:46.456364Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 2594, "message_tokens": 2501, "message_utf8_bytes": 6628, "cache_tensor_bytes": null, "serialized_payload_bytes": 6628, "side_information_bytes": 0, "fetched_bytes": 3416, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26388.96712497808, "ttft_ms": 13.599249999970198, "decode_ms": 94.49212509207428, "receiver_prefill_ms": 307.1271670050919}, "total_wall_ms": 26807.720249984413, "ttft_ms": 13.599249999970198, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "687,975", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "d4bb17e378679eb3945a0771dfcc20f8f33d0e49709ec7a673e887bbb8ae5e1f", "output_hash": "ef19034975881d08c60ac17454ae5ab6c3e5ce2b4b41f46e1509aa90f9a2ee93"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c166988a6e15b5c51685e724", "attempt_id": "5fd698a0dced", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:30:46.899238Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 2546, "message_tokens": 2456, "message_utf8_bytes": 6561, "cache_tensor_bytes": null, "serialized_payload_bytes": 6561, "side_information_bytes": 0, "fetched_bytes": 3336, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26388.96712497808, "ttft_ms": 14.812375069595873, "decode_ms": 14.036459033377469, "receiver_prefill_ms": 350.62629100866616}, "total_wall_ms": 26772.52762496937, "ttft_ms": 14.812375069595873, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ba07ed4b8b14eb1ce0f9a81b46254e6a4e14be640581ee5194e1ea0263fb4456", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "651683015a43cdffe283d8fa", "attempt_id": "146c2c042470", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:30:47.299854Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 2621, "message_tokens": 2524, "message_utf8_bytes": 6663, "cache_tensor_bytes": null, "serialized_payload_bytes": 6663, "side_information_bytes": 0, "fetched_bytes": 3420, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26388.96712497808, "ttft_ms": 13.781875022687018, "decode_ms": 13.484333991073072, "receiver_prefill_ms": 530.2059589885175}, "total_wall_ms": 26950.507750036195, "ttft_ms": 13.781875022687018, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "1ddc927267ee9d5539c10042f8829843c53f6c888ab73f7ef106acc0f93f3242", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "18517989dcb4fe4e1bbb18ed", "attempt_id": "bab0f4b6ba9c", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:30:47.879647Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 2614, "message_tokens": 2524, "message_utf8_bytes": 6663, "cache_tensor_bytes": null, "serialized_payload_bytes": 6663, "side_information_bytes": 0, "fetched_bytes": 3420, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26388.96712497808, "ttft_ms": 13.971916981972754, "decode_ms": 82.0794579340145, "receiver_prefill_ms": 298.2644169824198}, "total_wall_ms": 26786.69970901683, "ttft_ms": 13.971916981972754, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "292855", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "1ddc927267ee9d5539c10042f8829843c53f6c888ab73f7ef106acc0f93f3242", "output_hash": "4d2a03221fc71f99807c30668f5eb1a1311643cf5a57c7f7a36cc0c09df5f518"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "accccdf6b0f71476496c1ca0", "attempt_id": "89ad7c4edd60", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:30:48.294574Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 2603, "message_tokens": 2503, "message_utf8_bytes": 6605, "cache_tensor_bytes": null, "serialized_payload_bytes": 6605, "side_information_bytes": 0, "fetched_bytes": 3384, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26388.96712497808, "ttft_ms": 13.615000061690807, "decode_ms": 95.5932920332998, "receiver_prefill_ms": 287.9390419693664}, "total_wall_ms": 26789.543834049255, "ttft_ms": 13.615000061690807, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Talostmel-19", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "d721b0a442872f86f11822da780e30a5b0e1bbfe116d4483abc7173a5ea6eb4e", "output_hash": "d36ce632b9874676c785373c6928c2f5d2c3ef3b6f327220429c713e0d2e5490"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "dfa40c3df14984f6dfbfc759", "attempt_id": "965879ae0efd", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:30:48.712928Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 2615, "message_tokens": 2522, "message_utf8_bytes": 6694, "cache_tensor_bytes": null, "serialized_payload_bytes": 6694, "side_information_bytes": 0, "fetched_bytes": 3448, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26388.96712497808, "ttft_ms": 13.640958000905812, "decode_ms": 96.40233300160617, "receiver_prefill_ms": 300.0372499227524}, "total_wall_ms": 26802.61833302211, "ttft_ms": 13.640958000905812, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "2341350", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "15357f6c66639551691b74dea44989baa403c45954ec0d1ca978c56217ae5fe0", "output_hash": "4cbdc8d4bde64906434162984967a1eb27b9a045e9b2dca8ccadca9836a480d7"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "e5d1390ed9688ba0b2d58690", "attempt_id": "d86138fadd52", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:30:49.145104Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 2592, "message_tokens": 2498, "message_utf8_bytes": 6651, "cache_tensor_bytes": null, "serialized_payload_bytes": 6651, "side_information_bytes": 0, "fetched_bytes": 3397, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26388.96712497808, "ttft_ms": 13.876750017516315, "decode_ms": 13.817957951687276, "receiver_prefill_ms": 323.28825001604855}, "total_wall_ms": 26744.25191595219, "ttft_ms": 13.876750017516315, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "3ef6b0489e9defd32b1f903826e5558af18f694967e5e652335b5314840a6df2", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c89dff034f85c859e3889c6c", "attempt_id": "6b58720ae64f", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:30:49.528627Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 2619, "message_tokens": 2524, "message_utf8_bytes": 6663, "cache_tensor_bytes": null, "serialized_payload_bytes": 6663, "side_information_bytes": 0, "fetched_bytes": 3420, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26388.96712497808, "ttft_ms": 13.88912508264184, "decode_ms": 13.854042044840753, "receiver_prefill_ms": 311.8051670026034}, "total_wall_ms": 26732.018083916046, "ttft_ms": 13.88912508264184, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "e59312c5b542ffcccd40c34da8936eca082244b7db4dba9e53c1d641518bb845", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "e1af61428d53ed6739322568", "attempt_id": "3a2dc4df3920", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:30:50.936919Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 237957120}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 30632, "cache_tensor_bytes": 237957120, "serialized_payload_bytes": 237957120, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 1047.6557079236954, "ttft_ms": 52.27395903784782, "decode_ms": 140.6419170089066, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1247.2787499427795, "ttft_ms": 52.27395903784782, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7,120,879", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "c013fa074e4007a6fc0636d19a80febc8aee860429794ef4bbbe381185afe02c"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "570f36de1c7c0164ddaf9ffc", "attempt_id": "e2261e3cee0a", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:30:51.153962Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 237957120}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 44, "message_tokens": null, "message_utf8_bytes": 30632, "cache_tensor_bytes": 237957120, "serialized_payload_bytes": 237957120, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 1047.6557079236954, "ttft_ms": 48.579333000816405, "decode_ms": 150.86141601204872, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1251.6609999584034, "ttft_ms": 48.579333000816405, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "GX-54Q-4853", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "8a4b595bf859d0e547dcbb91a1485552981f1f40252e9575acc6751bff785bea"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "437abf6b717eb2155c7752c5", "attempt_id": "eb3fafbc5465", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:30:51.376084Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 237957120}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 51, "message_tokens": null, "message_utf8_bytes": 30632, "cache_tensor_bytes": 237957120, "serialized_payload_bytes": 237957120, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 1047.6557079236954, "ttft_ms": 91.02883306331933, "decode_ms": 105.66420806571841, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1248.8432909594849, "ttft_ms": 91.02883306331933, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "6041915", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "5f8f3414199be3bad75f36df45e3759125ec38c5a7eccea16448f93268c5fa92"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "740e45986dd4a2b1bb4c750e", "attempt_id": "d1833e62899d", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:30:51.595146Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 237957120}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 44, "message_tokens": null, "message_utf8_bytes": 30632, "cache_tensor_bytes": 237957120, "serialized_payload_bytes": 237957120, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 1047.6557079236954, "ttft_ms": 50.18000002019107, "decode_ms": 15.201958012767136, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1117.8218749118969, "ttft_ms": 50.18000002019107, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b9f4c8cae0c2363bd2a42da2", "attempt_id": "0f3117e304f4", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:30:51.683155Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 237957120}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 54, "message_tokens": null, "message_utf8_bytes": 30632, "cache_tensor_bytes": 237957120, "serialized_payload_bytes": 237957120, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 1047.6557079236954, "ttft_ms": 97.44029096327722, "decode_ms": 91.51095896959305, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1241.2401239853352, "ttft_ms": 97.44029096327722, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Rifel-16", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "34afeff5cc7f06df0544f826036c046cb8534ac4054132ee05dccf975102a81b"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "9d3c8b4c8cddfa77c10308f1", "attempt_id": "52fb359c418e", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:30:51.894516Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 237957120}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 30632, "cache_tensor_bytes": 237957120, "serialized_payload_bytes": 237957120, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 1047.6557079236954, "ttft_ms": 49.681207980029285, "decode_ms": 105.38004199042916, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1207.3754579760134, "ttft_ms": 49.681207980029285, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8066734", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "167c75fdd6a41fca64228797ab8f3a1a3f994475246df0bffb2e060d5d2e4ff0"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "899f2e78900aeaca31d086a7", "attempt_id": "9109b5c33c25", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:30:52.072108Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 237957120}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 48, "message_tokens": null, "message_utf8_bytes": 30632, "cache_tensor_bytes": 237957120, "serialized_payload_bytes": 237957120, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 1047.6557079236954, "ttft_ms": 49.68554200604558, "decode_ms": 15.216332976706326, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1117.1349169453606, "ttft_ms": 49.68554200604558, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "1bf53c172495f7685ea481fb", "attempt_id": "7bfbd8389173", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:30:52.160007Z", "task_family": "synthetic", "document_id": "syn_03022_8000", "document_hash": "c14b1263e320963acc2c01abd7735e9befb8a39cf89cb083c4d95987e8d38eee", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03022_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 237957120}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5828, "query_tokens": 49, "message_tokens": null, "message_utf8_bytes": 30632, "cache_tensor_bytes": 237957120, "serialized_payload_bytes": 237957120, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 1047.6557079236954, "ttft_ms": 47.359707998111844, "decode_ms": 15.107375103980303, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1114.7937079658732, "ttft_ms": 47.359707998111844, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Selbanwen-10", "bogus_amount": "6,245,875", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03022_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "40c04ffc8e76879bfeb124c9", "attempt_id": "0b039921526f", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:30:53.163218Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 2498, "message_tokens": 2414, "message_utf8_bytes": 8564, "cache_tensor_bytes": null, "serialized_payload_bytes": 8564, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.030542025342583656, "ttft_ms": 13.566375011578202, "decode_ms": 94.80612503830343, "receiver_prefill_ms": 282.27895905729383}, "total_wall_ms": 395.0876259477809, "ttft_ms": 13.566375011578202, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "6009480", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "4e9184df87a0cd5d465ab41f2efd29c7b77745ad3e4835c197306f8c37079e93", "output_hash": "925ebdb42a0311d11ea73517fd780f133bed1e6fad2f9f6d66408fe5dba57a45"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "587a2b30b13426e98aaf904d", "attempt_id": "f244c00581a3", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:30:53.576136Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 2502, "message_tokens": 2414, "message_utf8_bytes": 8564, "cache_tensor_bytes": null, "serialized_payload_bytes": 8564, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.030542025342583656, "ttft_ms": 13.44966609030962, "decode_ms": 134.54345799982548, "receiver_prefill_ms": 271.2700000265613}, "total_wall_ms": 425.94708397518843, "ttft_ms": 13.44966609030962, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "CT-20T-7883", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "4e9184df87a0cd5d465ab41f2efd29c7b77745ad3e4835c197306f8c37079e93", "output_hash": "dd2e4bce724b0497351a41dbf71da774970bfcbe2553b73571a02bc7e8fa0e53"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c3d53286ed5d622b0bfecbcd", "attempt_id": "aaf4f233720a", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:30:54.016419Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 2509, "message_tokens": 2414, "message_utf8_bytes": 8564, "cache_tensor_bytes": null, "serialized_payload_bytes": 8564, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.030542025342583656, "ttft_ms": 13.807207928039134, "decode_ms": 95.26216692756861, "receiver_prefill_ms": 271.1039580171928}, "total_wall_ms": 384.1880839318037, "ttft_ms": 13.807207928039134, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "6786800", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "4e9184df87a0cd5d465ab41f2efd29c7b77745ad3e4835c197306f8c37079e93", "output_hash": "33cee35795111ffd7720bcc954c699ed0ef5e51fe4ea1091cab2d69cd6de104f"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "87d33ef10cd19995de0abb69", "attempt_id": "da307c6284db", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:30:54.415688Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 2500, "message_tokens": 2414, "message_utf8_bytes": 8564, "cache_tensor_bytes": null, "serialized_payload_bytes": 8564, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.030542025342583656, "ttft_ms": 14.262999990023673, "decode_ms": 124.98570792376995, "receiver_prefill_ms": 271.1082500172779}, "total_wall_ms": 414.1628759680316, "ttft_ms": 14.262999990023673, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8,810,500", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "4e9184df87a0cd5d465ab41f2efd29c7b77745ad3e4835c197306f8c37079e93", "output_hash": "7810d312d2690f45bb13651b3c1cf9a2098b9e9d043463aaccc81157fad4a26e"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "dabc91782f625c51f85143e8", "attempt_id": "bceff53d1ab0", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:30:54.843795Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 2510, "message_tokens": 2414, "message_utf8_bytes": 8564, "cache_tensor_bytes": null, "serialized_payload_bytes": 8564, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.030542025342583656, "ttft_ms": 13.756499974988401, "decode_ms": 81.98524999897927, "receiver_prefill_ms": 274.6542498935014}, "total_wall_ms": 379.3360011186451, "ttft_ms": 13.756499974988401, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Quidra-11", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "4e9184df87a0cd5d465ab41f2efd29c7b77745ad3e4835c197306f8c37079e93", "output_hash": "41698c622e83fd669a8d097b3f00cb4197b0a66bac5aeca05b6224ada8b9dd5d"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a74560e348184248cbc606d7", "attempt_id": "3e4aa0e32475", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:30:55.238005Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 2503, "message_tokens": 2414, "message_utf8_bytes": 8564, "cache_tensor_bytes": null, "serialized_payload_bytes": 8564, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.030542025342583656, "ttft_ms": 13.611166970804334, "decode_ms": 135.92820905614644, "receiver_prefill_ms": 273.4401250490919}, "total_wall_ms": 426.7311670118943, "ttft_ms": 13.611166970804334, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "42,050,330", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "4e9184df87a0cd5d465ab41f2efd29c7b77745ad3e4835c197306f8c37079e93", "output_hash": "6306e3124ddb9d73a354dce7add522b5e4bb245962bf0f7c09537621a96c30b5"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "8544b8a2f8072cc1982929ab", "attempt_id": "799152cb7a5c", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:30:55.678748Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 2504, "message_tokens": 2414, "message_utf8_bytes": 8564, "cache_tensor_bytes": null, "serialized_payload_bytes": 8564, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.030542025342583656, "ttft_ms": 13.363540987484157, "decode_ms": 13.921667006798089, "receiver_prefill_ms": 270.2587080420926}, "total_wall_ms": 301.2395000550896, "ttft_ms": 13.363540987484157, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "4e9184df87a0cd5d465ab41f2efd29c7b77745ad3e4835c197306f8c37079e93", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "eb3a23d297cbe608c195b7b9", "attempt_id": "fb82ca04943f", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:30:55.993975Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 2505, "message_tokens": 2414, "message_utf8_bytes": 8564, "cache_tensor_bytes": null, "serialized_payload_bytes": 8564, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.030542025342583656, "ttft_ms": 13.556667021475732, "decode_ms": 13.639707933180034, "receiver_prefill_ms": 271.27550006844103}, "total_wall_ms": 302.1440840093419, "ttft_ms": 13.556667021475732, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "4e9184df87a0cd5d465ab41f2efd29c7b77745ad3e4835c197306f8c37079e93", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "212c4ba92d939cf4495ee1d4", "attempt_id": "be6a3641dce3", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:31:10.416782Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 1122, "message_tokens": 1037, "message_utf8_bytes": 2682, "cache_tensor_bytes": null, "serialized_payload_bytes": 2682, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14106.442834017798, "ttft_ms": 13.415291905403137, "decode_ms": 94.56825000233948, "receiver_prefill_ms": 122.01437493786216}, "total_wall_ms": 14338.359251036309, "ttft_ms": 13.415291905403137, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "6009480", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ae4d4d02d5d15b5daa692263d634313c88e96ac068a189d8c115e529c43ff39d", "output_hash": "925ebdb42a0311d11ea73517fd780f133bed1e6fad2f9f6d66408fe5dba57a45"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "944620e6b89d44c34fbe2bc2", "attempt_id": "f575452d8c67", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:31:10.663287Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 1126, "message_tokens": 1037, "message_utf8_bytes": 2682, "cache_tensor_bytes": null, "serialized_payload_bytes": 2682, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14106.442834017798, "ttft_ms": 13.482582988217473, "decode_ms": 134.18254198040813, "receiver_prefill_ms": 121.90004205331206}, "total_wall_ms": 14377.818958950229, "ttft_ms": 13.482582988217473, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "CT-20T-7883", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ae4d4d02d5d15b5daa692263d634313c88e96ac068a189d8c115e529c43ff39d", "output_hash": "dd2e4bce724b0497351a41dbf71da774970bfcbe2553b73571a02bc7e8fa0e53"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "681e424fd1720577dbef1cd1", "attempt_id": "02fa8ab49fd6", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:31:10.948778Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 1133, "message_tokens": 1037, "message_utf8_bytes": 2682, "cache_tensor_bytes": null, "serialized_payload_bytes": 2682, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14106.442834017798, "ttft_ms": 13.485042029060423, "decode_ms": 121.07954209204763, "receiver_prefill_ms": 121.89658300485462}, "total_wall_ms": 14364.784084027633, "ttft_ms": 13.485042029060423, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "6,786,800", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ae4d4d02d5d15b5daa692263d634313c88e96ac068a189d8c115e529c43ff39d", "output_hash": "396a4b6b618fb10a20ca84793c54e794cbfebea7547dd351971f763f30153ef9"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "dae5030a0f42d02d6921cbe3", "attempt_id": "b61c7ff3b0ec", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:31:11.221070Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 1124, "message_tokens": 1037, "message_utf8_bytes": 2682, "cache_tensor_bytes": null, "serialized_payload_bytes": 2682, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14106.442834017798, "ttft_ms": 13.426209101453424, "decode_ms": 94.41970800980926, "receiver_prefill_ms": 121.1685830494389}, "total_wall_ms": 14337.342917104252, "ttft_ms": 13.426209101453424, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7510570", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ae4d4d02d5d15b5daa692263d634313c88e96ac068a189d8c115e529c43ff39d", "output_hash": "410d7d10eb5e18b187a54949b6517bffab264255a8ac37e64c7fa5b57e44d9ed"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "6656ba040d1d285c18734a57", "attempt_id": "f68e0f5202ad", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:31:11.465875Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 1134, "message_tokens": 1037, "message_utf8_bytes": 2682, "cache_tensor_bytes": null, "serialized_payload_bytes": 2682, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14106.442834017798, "ttft_ms": 13.432958978228271, "decode_ms": 80.76499996241182, "receiver_prefill_ms": 121.36554205790162}, "total_wall_ms": 14323.872125940397, "ttft_ms": 13.432958978228271, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Quilun-16", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ae4d4d02d5d15b5daa692263d634313c88e96ac068a189d8c115e529c43ff39d", "output_hash": "34a7c0e11c93ba75eb4b20d3d3f5303b417856d89a232e8b16fef75c4336afc6"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "6dead528c8b7412e99109fdf", "attempt_id": "25efc85f102d", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:31:11.697413Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 1127, "message_tokens": 1037, "message_utf8_bytes": 2682, "cache_tensor_bytes": null, "serialized_payload_bytes": 2682, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14106.442834017798, "ttft_ms": 13.51770805194974, "decode_ms": 134.3944170512259, "receiver_prefill_ms": 120.95116602722555}, "total_wall_ms": 14377.248583943583, "ttft_ms": 13.51770805194974, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "32,577,719", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ae4d4d02d5d15b5daa692263d634313c88e96ac068a189d8c115e529c43ff39d", "output_hash": "bcb604e76cda7a500e75221be955eba37a37232877302c549bc895816d55ba73"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "3ad96b2ade930f877015f88e", "attempt_id": "b7585c3b86e9", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:31:11.982534Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 1128, "message_tokens": 1037, "message_utf8_bytes": 2682, "cache_tensor_bytes": null, "serialized_payload_bytes": 2682, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14106.442834017798, "ttft_ms": 13.517333078198135, "decode_ms": 13.72904097661376, "receiver_prefill_ms": 120.82354188896716}, "total_wall_ms": 14256.329125026241, "ttft_ms": 13.517333078198135, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ae4d4d02d5d15b5daa692263d634313c88e96ac068a189d8c115e529c43ff39d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "103a3b5dcce005e7de1a8cb5", "attempt_id": "c904952dd07e", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:31:12.146185Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 1129, "message_tokens": 1037, "message_utf8_bytes": 2682, "cache_tensor_bytes": null, "serialized_payload_bytes": 2682, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14106.442834017798, "ttft_ms": 13.342999969609082, "decode_ms": 13.240540982224047, "receiver_prefill_ms": 121.14233407191932}, "total_wall_ms": 14255.988417076878, "ttft_ms": 13.342999969609082, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ae4d4d02d5d15b5daa692263d634313c88e96ac068a189d8c115e529c43ff39d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "aafb5a397358863c106c1bb9", "attempt_id": "2dee65fe2f36", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:31:24.824881Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 13}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 584, "message_tokens": 495, "message_utf8_bytes": 1363, "cache_tensor_bytes": null, "serialized_payload_bytes": 1363, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 12515.18854207825, "ttft_ms": 13.558125006966293, "decode_ms": 13.425957993604243, "receiver_prefill_ms": 69.91595798172057}, "total_wall_ms": 12613.347416976467, "ttft_ms": 13.558125006966293, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "b0b7513f495aab95aaf60ea9bb746e048270259dd15a42b334b89a5585522417", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "cc677eaae4e7a2e0d2490ff5", "attempt_id": "20f334bc6a91", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:31:24.937918Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 13}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 588, "message_tokens": 495, "message_utf8_bytes": 1363, "cache_tensor_bytes": null, "serialized_payload_bytes": 1363, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 12515.18854207825, "ttft_ms": 13.247665949165821, "decode_ms": 13.465916970744729, "receiver_prefill_ms": 69.78424999397248}, "total_wall_ms": 12612.862750072964, "ttft_ms": 13.247665949165821, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "b0b7513f495aab95aaf60ea9bb746e048270259dd15a42b334b89a5585522417", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "181a9728291b0da178001020", "attempt_id": "e586c683b6a7", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:31:25.049632Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 13}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 595, "message_tokens": 495, "message_utf8_bytes": 1363, "cache_tensor_bytes": null, "serialized_payload_bytes": 1363, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 12515.18854207825, "ttft_ms": 13.63187504466623, "decode_ms": 13.1967089837417, "receiver_prefill_ms": 69.63366596028209}, "total_wall_ms": 12612.848667078651, "ttft_ms": 13.63187504466623, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "b0b7513f495aab95aaf60ea9bb746e048270259dd15a42b334b89a5585522417", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c5f74a20b6325bb5cb9f1a03", "attempt_id": "804a9110ce80", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:31:25.161229Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 13}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 586, "message_tokens": 495, "message_utf8_bytes": 1363, "cache_tensor_bytes": null, "serialized_payload_bytes": 1363, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 12515.18854207825, "ttft_ms": 13.486499898135662, "decode_ms": 13.725540949963033, "receiver_prefill_ms": 69.51070809736848}, "total_wall_ms": 12613.076084060594, "ttft_ms": 13.486499898135662, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "b0b7513f495aab95aaf60ea9bb746e048270259dd15a42b334b89a5585522417", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "d1eff5b610e2a18ad9ae6c69", "attempt_id": "81d6314f8fe4", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:31:25.273063Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 13}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 596, "message_tokens": 495, "message_utf8_bytes": 1363, "cache_tensor_bytes": null, "serialized_payload_bytes": 1363, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 12515.18854207825, "ttft_ms": 13.119707931764424, "decode_ms": 79.7553330194205, "receiver_prefill_ms": 69.46679204702377}, "total_wall_ms": 12678.766376106068, "ttft_ms": 13.119707931764424, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Draard-12", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "b0b7513f495aab95aaf60ea9bb746e048270259dd15a42b334b89a5585522417", "output_hash": "c98ed1d4a10f5c2d70e15663c4ed0df74b420086785b671165017dfed27dee41"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "97981d54106a5f9841d26521", "attempt_id": "195923a5bdf5", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:31:25.450891Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 13}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 589, "message_tokens": 495, "message_utf8_bytes": 1363, "cache_tensor_bytes": null, "serialized_payload_bytes": 1363, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 12515.18854207825, "ttft_ms": 13.343999977223575, "decode_ms": 93.33691699430346, "receiver_prefill_ms": 69.34654095675796}, "total_wall_ms": 12692.408458096907, "ttft_ms": 13.343999977223575, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8066734", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "b0b7513f495aab95aaf60ea9bb746e048270259dd15a42b334b89a5585522417", "output_hash": "167c75fdd6a41fca64228797ab8f3a1a3f994475246df0bffb2e060d5d2e4ff0"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "1407ec4021733dbee950003a", "attempt_id": "26602c8e8ee1", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:31:25.641953Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 13}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 590, "message_tokens": 495, "message_utf8_bytes": 1363, "cache_tensor_bytes": null, "serialized_payload_bytes": 1363, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 12515.18854207825, "ttft_ms": 13.271415955387056, "decode_ms": 13.361167046241462, "receiver_prefill_ms": 69.67112503480166}, "total_wall_ms": 12612.690209178254, "ttft_ms": 13.271415955387056, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "b0b7513f495aab95aaf60ea9bb746e048270259dd15a42b334b89a5585522417", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "3ff747f08168df1f8de20f59", "attempt_id": "1bfce3ccec29", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:31:25.753596Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 13}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 591, "message_tokens": 495, "message_utf8_bytes": 1363, "cache_tensor_bytes": null, "serialized_payload_bytes": 1363, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 12515.18854207825, "ttft_ms": 13.252458069473505, "decode_ms": 13.157709036022425, "receiver_prefill_ms": 69.44150000344962}, "total_wall_ms": 12612.151834066026, "ttft_ms": 13.252458069473505, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "b0b7513f495aab95aaf60ea9bb746e048270259dd15a42b334b89a5585522417", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "2e28fcf23d2c72b1304c359c", "attempt_id": "b1e25fb2d292", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:31:38.379802Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 1397, "message_tokens": 1309, "message_utf8_bytes": 3271, "cache_tensor_bytes": null, "serialized_payload_bytes": 3271, "side_information_bytes": 0, "fetched_bytes": 1790, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 12515.177167020738, "ttft_ms": 13.57225002720952, "decode_ms": 94.59033398889005, "receiver_prefill_ms": 148.48066691774875}, "total_wall_ms": 12773.999167024158, "ttft_ms": 13.57225002720952, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "6009480", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "4a88b73a8ccb18945474231c18c56dd9cf1e6060f34547e483e5ea652a087815", "output_hash": "925ebdb42a0311d11ea73517fd780f133bed1e6fad2f9f6d66408fe5dba57a45"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "e5412bb9358f4c35b80262f9", "attempt_id": "4d7ac4198314", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:31:38.654072Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 1368, "message_tokens": 1276, "message_utf8_bytes": 3188, "cache_tensor_bytes": null, "serialized_payload_bytes": 3188, "side_information_bytes": 0, "fetched_bytes": 1710, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 12515.177167020738, "ttft_ms": 13.374125002883375, "decode_ms": 13.208957971073687, "receiver_prefill_ms": 144.81504203286022}, "total_wall_ms": 12688.624042086303, "ttft_ms": 13.374125002883375, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ffff6ea4abd0ba6ff6525a96e7c48600de08addd5f854fd3b688b2f3cd1f1e08", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b98b3dc48b11c5d30c2d6681", "attempt_id": "ae7e33fb0a47", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:31:38.841385Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 1408, "message_tokens": 1309, "message_utf8_bytes": 3271, "cache_tensor_bytes": null, "serialized_payload_bytes": 3271, "side_information_bytes": 0, "fetched_bytes": 1790, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 12515.177167020738, "ttft_ms": 13.490332989022136, "decode_ms": 13.323208084329963, "receiver_prefill_ms": 148.87733396608382}, "total_wall_ms": 12692.991917021573, "ttft_ms": 13.490332989022136, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "4a88b73a8ccb18945474231c18c56dd9cf1e6060f34547e483e5ea652a087815", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "55e25b9a4020b6012f4b9ce7", "attempt_id": "20f7f3c95157", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:31:39.034292Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 1399, "message_tokens": 1309, "message_utf8_bytes": 3271, "cache_tensor_bytes": null, "serialized_payload_bytes": 3271, "side_information_bytes": 0, "fetched_bytes": 1790, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 12515.177167020738, "ttft_ms": 13.373833033256233, "decode_ms": 94.83158309012651, "receiver_prefill_ms": 148.91783299390227}, "total_wall_ms": 12774.377208086662, "ttft_ms": 13.373833033256233, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7510570", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "923289d96e921e46621d25ce69d26d0f453ecf0b6e66e67bfc18366cc6d76587", "output_hash": "410d7d10eb5e18b187a54949b6517bffab264255a8ac37e64c7fa5b57e44d9ed"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "931267ce168e3611bd809c9b", "attempt_id": "9b6a3a90e659", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:31:39.307334Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 1405, "message_tokens": 1305, "message_utf8_bytes": 3270, "cache_tensor_bytes": null, "serialized_payload_bytes": 3270, "side_information_bytes": 0, "fetched_bytes": 1787, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 12515.177167020738, "ttft_ms": 13.56566697359085, "decode_ms": 80.08491701912135, "receiver_prefill_ms": 148.89662503264844}, "total_wall_ms": 12759.98579210136, "ttft_ms": 13.56566697359085, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Draard-12", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "e2b3037d8908da06298f667c318fdee9bfa5c1b9e0d87cba8cdbc052ed0bee9a", "output_hash": "c98ed1d4a10f5c2d70e15663c4ed0df74b420086785b671165017dfed27dee41"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "82a8599d60aa70c2e5515b6f", "attempt_id": "dfbff9d583b2", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:31:39.566815Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 1402, "message_tokens": 1309, "message_utf8_bytes": 3271, "cache_tensor_bytes": null, "serialized_payload_bytes": 3271, "side_information_bytes": 0, "fetched_bytes": 1790, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 12515.177167020738, "ttft_ms": 13.415834051556885, "decode_ms": 108.37150004226714, "receiver_prefill_ms": 148.01033306866884}, "total_wall_ms": 12787.11758297868, "ttft_ms": 13.415834051556885, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "26748811", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "4a88b73a8ccb18945474231c18c56dd9cf1e6060f34547e483e5ea652a087815", "output_hash": "d49d87bb3b997d06c0986aed2e51ae3910ffbbc442a2f12cfe5239c2da5e9353"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "27cb3091caf3805f9223f66a", "attempt_id": "d7d6b5a93946", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:31:39.852696Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 1391, "message_tokens": 1297, "message_utf8_bytes": 3259, "cache_tensor_bytes": null, "serialized_payload_bytes": 3259, "side_information_bytes": 0, "fetched_bytes": 1771, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 12515.177167020738, "ttft_ms": 13.529125018976629, "decode_ms": 13.386375037953258, "receiver_prefill_ms": 148.83320801891387}, "total_wall_ms": 12693.106333026662, "ttft_ms": 13.529125018976629, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "308afd5f5d91e3e5623fb013b660fe534657ff7706da0041c6a822c06c814e97", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "8caf002abd225f4f55990119", "attempt_id": "a647fc9864e3", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:31:40.045379Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 1404, "message_tokens": 1309, "message_utf8_bytes": 3271, "cache_tensor_bytes": null, "serialized_payload_bytes": 3271, "side_information_bytes": 0, "fetched_bytes": 1790, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 12515.177167020738, "ttft_ms": 13.566749985329807, "decode_ms": 13.344415929168463, "receiver_prefill_ms": 148.42520805541426}, "total_wall_ms": 12692.65387498308, "ttft_ms": 13.566749985329807, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "da7ea632e2f2b1db4e344cab53e795cea56f3fad706e86cc719fcc96ffc0e31d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "55e09c3560e9084a9c0150b4", "attempt_id": "dc3af018ce62", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:31:40.532560Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 91865088}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 42, "message_tokens": null, "message_utf8_bytes": 8538, "cache_tensor_bytes": 91865088, "serialized_payload_bytes": 91865088, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 295.14358402229846, "ttft_ms": 34.026875044219196, "decode_ms": 98.1809999793768, "receiver_prefill_ms": 0.0}, "total_wall_ms": 431.9698759354651, "ttft_ms": 34.026875044219196, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8066734", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "167c75fdd6a41fca64228797ab8f3a1a3f994475246df0bffb2e060d5d2e4ff0"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c248e8f65d3f8df84c1d68b0", "attempt_id": "d535d06a9b33", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:31:40.683414Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 91865088}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 46, "message_tokens": null, "message_utf8_bytes": 8538, "cache_tensor_bytes": 91865088, "serialized_payload_bytes": 91865088, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 295.14358402229846, "ttft_ms": 28.9333340479061, "decode_ms": 140.4450419358909, "receiver_prefill_ms": 0.0}, "total_wall_ms": 468.22804294060916, "ttft_ms": 28.9333340479061, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "CT-20T-7883", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "dd2e4bce724b0497351a41dbf71da774970bfcbe2553b73571a02bc7e8fa0e53"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "d7aecbf8fbca4d67c846acaf", "attempt_id": "67c0b0c9fb2a", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:31:40.870272Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 91865088}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 53, "message_tokens": null, "message_utf8_bytes": 8538, "cache_tensor_bytes": 91865088, "serialized_payload_bytes": 91865088, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 295.14358402229846, "ttft_ms": 29.38929107040167, "decode_ms": 97.74766699410975, "receiver_prefill_ms": 0.0}, "total_wall_ms": 426.3515010243282, "ttft_ms": 29.38929107040167, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "6786800", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "33cee35795111ffd7720bcc954c699ed0ef5e51fe4ea1091cab2d69cd6de104f"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "54309fb3c9e431b014616f64", "attempt_id": "d78809cd372b", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:31:41.015819Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 91865088}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 44, "message_tokens": null, "message_utf8_bytes": 8538, "cache_tensor_bytes": 91865088, "serialized_payload_bytes": 91865088, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 295.14358402229846, "ttft_ms": 28.124709031544626, "decode_ms": 125.67083304747939, "receiver_prefill_ms": 0.0}, "total_wall_ms": 453.0987510224804, "ttft_ms": 28.124709031544626, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4,448,780", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "16bc7be434494b015249be21e11fa0b76d9a1d0019bef3e4aa915ad0774d29d1"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "8876250356a1a122bc187aea", "attempt_id": "c8c5f5ace8b9", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:31:41.188215Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 91865088}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 54, "message_tokens": null, "message_utf8_bytes": 8538, "cache_tensor_bytes": 91865088, "serialized_payload_bytes": 91865088, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 295.14358402229846, "ttft_ms": 29.017874971032143, "decode_ms": 84.32183298282325, "receiver_prefill_ms": 0.0}, "total_wall_ms": 412.3372919857502, "ttft_ms": 29.017874971032143, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Draard-12", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "c98ed1d4a10f5c2d70e15663c4ed0df74b420086785b671165017dfed27dee41"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "ba1101631fb0743c4c5946b2", "attempt_id": "49782b1f6476", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:31:41.319314Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 91865088}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 8538, "cache_tensor_bytes": 91865088, "serialized_payload_bytes": 91865088, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 295.14358402229846, "ttft_ms": 29.218041920103133, "decode_ms": 139.21016594395041, "receiver_prefill_ms": 0.0}, "total_wall_ms": 467.72679209243506, "ttft_ms": 29.218041920103133, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "31,964,570", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "439c7a5ac9d6bc24375211ea55b790b894a6911b8f28b92268fc9325d62aecce"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "04452384f3dcd11c6e19765d", "attempt_id": "232081c701b7", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:31:41.506416Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 91865088}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 48, "message_tokens": null, "message_utf8_bytes": 8538, "cache_tensor_bytes": 91865088, "serialized_payload_bytes": 91865088, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 295.14358402229846, "ttft_ms": 28.19379197899252, "decode_ms": 13.820708030834794, "receiver_prefill_ms": 0.0}, "total_wall_ms": 341.1353341070935, "ttft_ms": 28.19379197899252, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "76993531052a5587b47b42df", "attempt_id": "a62869141be0", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:31:41.566873Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 91865088}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 49, "message_tokens": null, "message_utf8_bytes": 8538, "cache_tensor_bytes": 91865088, "serialized_payload_bytes": 91865088, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 295.14358402229846, "ttft_ms": 29.5503749512136, "decode_ms": 13.893625000491738, "receiver_prefill_ms": 0.0}, "total_wall_ms": 342.7290430990979, "ttft_ms": 29.5503749512136, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a579979c20361ab60cbc7ca0", "attempt_id": "5e25257ae815", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:31:41.628563Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 2612, "message_tokens": 2528, "message_utf8_bytes": 9068, "cache_tensor_bytes": null, "serialized_payload_bytes": 9068, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.023875036276876926, "ttft_ms": 13.653250061906874, "decode_ms": 121.35220807977021, "receiver_prefill_ms": 283.6981250438839}, "total_wall_ms": 422.5835420656949, "ttft_ms": 13.653250061906874, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5,453,600", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "c06d757517ca8904b6f50914a780c15f35e8e0ba06e49f03c985c0581d1dceb3", "output_hash": "e6a2b25d3f99c6c8c339dc11738104244c59975cc325ebc53bc602e40b327216"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "43430ca106f4c77cdc38a804", "attempt_id": "2c6f30cb5865", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:31:42.065480Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 2616, "message_tokens": 2528, "message_utf8_bytes": 9068, "cache_tensor_bytes": null, "serialized_payload_bytes": 9068, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.023875036276876926, "ttft_ms": 13.439749949611723, "decode_ms": 134.8626249236986, "receiver_prefill_ms": 282.55808306857944}, "total_wall_ms": 434.8005000501871, "ttft_ms": 13.439749949611723, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "CT-20T-7883", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "c06d757517ca8904b6f50914a780c15f35e8e0ba06e49f03c985c0581d1dceb3", "output_hash": "dd2e4bce724b0497351a41dbf71da774970bfcbe2553b73571a02bc7e8fa0e53"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "4405f540d43eefd4de7c2e3f", "attempt_id": "327ff18f3f80", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:31:42.514326Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 2623, "message_tokens": 2528, "message_utf8_bytes": 9068, "cache_tensor_bytes": null, "serialized_payload_bytes": 9068, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.023875036276876926, "ttft_ms": 13.55433405842632, "decode_ms": 94.01562495622784, "receiver_prefill_ms": 283.1187079427764}, "total_wall_ms": 394.6062079630792, "ttft_ms": 13.55433405842632, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "6786800", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "c06d757517ca8904b6f50914a780c15f35e8e0ba06e49f03c985c0581d1dceb3", "output_hash": "33cee35795111ffd7720bcc954c699ed0ef5e51fe4ea1091cab2d69cd6de104f"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "00ff4590225e33950b1eafb5", "attempt_id": "28962261fe14", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:31:42.924073Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 2614, "message_tokens": 2528, "message_utf8_bytes": 9068, "cache_tensor_bytes": null, "serialized_payload_bytes": 9068, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.023875036276876926, "ttft_ms": 13.53074994403869, "decode_ms": 122.08979192655534, "receiver_prefill_ms": 283.10379199683666}, "total_wall_ms": 422.6961249951273, "ttft_ms": 13.53074994403869, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8,810,500", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "c06d757517ca8904b6f50914a780c15f35e8e0ba06e49f03c985c0581d1dceb3", "output_hash": "7810d312d2690f45bb13651b3c1cf9a2098b9e9d043463aaccc81157fad4a26e"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "8c1936812d1717cb77bb9593", "attempt_id": "a195f3ae066f", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:31:43.360940Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 2624, "message_tokens": 2528, "message_utf8_bytes": 9068, "cache_tensor_bytes": null, "serialized_payload_bytes": 9068, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.023875036276876926, "ttft_ms": 13.666000100784004, "decode_ms": 81.60508295986801, "receiver_prefill_ms": 283.1953330896795}, "total_wall_ms": 382.3897500988096, "ttft_ms": 13.666000100784004, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Quidra-11", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "c06d757517ca8904b6f50914a780c15f35e8e0ba06e49f03c985c0581d1dceb3", "output_hash": "41698c622e83fd669a8d097b3f00cb4197b0a66bac5aeca05b6224ada8b9dd5d"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "6f02a6b0e9290601a8b7ee07", "attempt_id": "ee0547ff2f4f", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:31:43.757494Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 2617, "message_tokens": 2528, "message_utf8_bytes": 9068, "cache_tensor_bytes": null, "serialized_payload_bytes": 9068, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.023875036276876926, "ttft_ms": 13.743832940235734, "decode_ms": 135.3820830117911, "receiver_prefill_ms": 293.2795420056209}, "total_wall_ms": 446.3246250525117, "ttft_ms": 13.743832940235734, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "40,045,313", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "c06d757517ca8904b6f50914a780c15f35e8e0ba06e49f03c985c0581d1dceb3", "output_hash": "e4df8dcf50c0cc11dacc1076e0717d84e5c0de040065c7bcdf59c1723f02ac29"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "41fe49b2eee4f73d8ab0d446", "attempt_id": "132b6a26a5d2", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:31:44.218414Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 2618, "message_tokens": 2528, "message_utf8_bytes": 9068, "cache_tensor_bytes": null, "serialized_payload_bytes": 9068, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.023875036276876926, "ttft_ms": 13.723165960982442, "decode_ms": 13.57287506107241, "receiver_prefill_ms": 283.7702500401065}, "total_wall_ms": 315.027084085159, "ttft_ms": 13.723165960982442, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "c06d757517ca8904b6f50914a780c15f35e8e0ba06e49f03c985c0581d1dceb3", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "1fb227206e36ddc7b6e6981b", "attempt_id": "e2de69a26b1d", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:31:44.547905Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 2619, "message_tokens": 2528, "message_utf8_bytes": 9068, "cache_tensor_bytes": null, "serialized_payload_bytes": 9068, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.023875036276876926, "ttft_ms": 13.76029197126627, "decode_ms": 13.69058305863291, "receiver_prefill_ms": 292.853832943365}, "total_wall_ms": 324.09112504683435, "ttft_ms": 13.76029197126627, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "c06d757517ca8904b6f50914a780c15f35e8e0ba06e49f03c985c0581d1dceb3", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "ca62bdceecb33759a381b20e", "attempt_id": "d627a5ed989b", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:32:08.197901Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1711}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 1820, "message_tokens": 1735, "message_utf8_bytes": 5015, "cache_tensor_bytes": null, "serialized_payload_bytes": 5015, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 23311.22629204765, "ttft_ms": 13.842749991454184, "decode_ms": 94.3367499858141, "receiver_prefill_ms": 194.15091595146805}, "total_wall_ms": 23617.153001134284, "ttft_ms": 13.842749991454184, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "6009480", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "0136aaee0b0c82e32b67ba5041e0cf3293a498bd9a294b0841d9787366b3680b", "output_hash": "925ebdb42a0311d11ea73517fd780f133bed1e6fad2f9f6d66408fe5dba57a45"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "461c10f16fe4d9d6ec66d7e1", "attempt_id": "360c688e0d97", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:32:08.518354Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1711}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 1824, "message_tokens": 1735, "message_utf8_bytes": 5015, "cache_tensor_bytes": null, "serialized_payload_bytes": 5015, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 23311.22629204765, "ttft_ms": 13.498624903149903, "decode_ms": 134.37291607260704, "receiver_prefill_ms": 194.57941700238734}, "total_wall_ms": 23656.23970911838, "ttft_ms": 13.498624903149903, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "CT-20T-7883", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "0136aaee0b0c82e32b67ba5041e0cf3293a498bd9a294b0841d9787366b3680b", "output_hash": "dd2e4bce724b0497351a41dbf71da774970bfcbe2553b73571a02bc7e8fa0e53"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "6d5043d6d7fbf2239beec5dd", "attempt_id": "bf5c85947a74", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:32:08.877918Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1711}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 1831, "message_tokens": 1735, "message_utf8_bytes": 5015, "cache_tensor_bytes": null, "serialized_payload_bytes": 5015, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 23311.22629204765, "ttft_ms": 13.351083034649491, "decode_ms": 94.77112500462681, "receiver_prefill_ms": 195.53579099010676}, "total_wall_ms": 23617.5482920371, "ttft_ms": 13.351083034649491, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5821605", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "0136aaee0b0c82e32b67ba5041e0cf3293a498bd9a294b0841d9787366b3680b", "output_hash": "66b2a9269bfd2b9ccbe3cf3b211a218de7702bc720ef7b486355c83813fdaed3"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "1a80d50aef8fa81041938b79", "attempt_id": "014ab596ad44", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:32:09.198521Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1711}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 1822, "message_tokens": 1735, "message_utf8_bytes": 5015, "cache_tensor_bytes": null, "serialized_payload_bytes": 5015, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 23311.22629204765, "ttft_ms": 13.587750028818846, "decode_ms": 95.77037510462105, "receiver_prefill_ms": 192.7806659368798}, "total_wall_ms": 23616.11266701948, "ttft_ms": 13.587750028818846, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7510570", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "0136aaee0b0c82e32b67ba5041e0cf3293a498bd9a294b0841d9787366b3680b", "output_hash": "410d7d10eb5e18b187a54949b6517bffab264255a8ac37e64c7fa5b57e44d9ed"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "99a031e24a87ef6847a85c0d", "attempt_id": "0b31ea4ee052", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:32:09.517813Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1711}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 1832, "message_tokens": 1735, "message_utf8_bytes": 5015, "cache_tensor_bytes": null, "serialized_payload_bytes": 5015, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 23311.22629204765, "ttft_ms": 13.879625010304153, "decode_ms": 81.87324996106327, "receiver_prefill_ms": 194.44316590670496}, "total_wall_ms": 23604.112416971475, "ttft_ms": 13.879625010304153, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Draard-12", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "0136aaee0b0c82e32b67ba5041e0cf3293a498bd9a294b0841d9787366b3680b", "output_hash": "c98ed1d4a10f5c2d70e15663c4ed0df74b420086785b671165017dfed27dee41"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a77e1f971cdbbb02fe196146", "attempt_id": "123eeb0359ea", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:32:09.824747Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1711}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 1825, "message_tokens": 1735, "message_utf8_bytes": 5015, "cache_tensor_bytes": null, "serialized_payload_bytes": 5015, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 23311.22629204765, "ttft_ms": 13.925459003075957, "decode_ms": 94.652249943465, "receiver_prefill_ms": 199.0933750057593}, "total_wall_ms": 23621.620875084773, "ttft_ms": 13.925459003075957, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "3484480", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "0136aaee0b0c82e32b67ba5041e0cf3293a498bd9a294b0841d9787366b3680b", "output_hash": "63cfa217c09892c2da9620cdc79438d18a8f82a38c8adf7ef814d6aad1974f6e"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "8ed56f57f11c22ed38f48d1b", "attempt_id": "2e9de55b5779", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:32:10.150097Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1711}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 1826, "message_tokens": 1735, "message_utf8_bytes": 5015, "cache_tensor_bytes": null, "serialized_payload_bytes": 5015, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 23311.22629204765, "ttft_ms": 14.270749990828335, "decode_ms": 14.437415986321867, "receiver_prefill_ms": 853.6735830130056}, "total_wall_ms": 24196.421584114432, "ttft_ms": 14.270749990828335, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "0136aaee0b0c82e32b67ba5041e0cf3293a498bd9a294b0841d9787366b3680b", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a1f2798d105cebbdf7c8dc74", "attempt_id": "e58929aaf0ec", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:32:11.051389Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1711}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 1827, "message_tokens": 1735, "message_utf8_bytes": 5015, "cache_tensor_bytes": null, "serialized_payload_bytes": 5015, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 23311.22629204765, "ttft_ms": 14.310833066701889, "decode_ms": 14.01895796880126, "receiver_prefill_ms": 1000.5094589432701}, "total_wall_ms": 24343.141708988696, "ttft_ms": 14.310833066701889, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "0136aaee0b0c82e32b67ba5041e0cf3293a498bd9a294b0841d9787366b3680b", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "31a14f912be37878c8241001", "attempt_id": "5934ad41cbac", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:32:39.432746Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 30}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 1238, "message_tokens": 1149, "message_utf8_bytes": 3043, "cache_tensor_bytes": null, "serialized_payload_bytes": 3043, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27333.36341707036, "ttft_ms": 14.396667014807463, "decode_ms": 121.00066698621958, "receiver_prefill_ms": 135.93758293427527}, "total_wall_ms": 27610.006709117442, "ttft_ms": 14.396667014807463, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "6,009,480", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "2dd52fe72ca47aacd0e383e165bc8f70c1b82d5f1667eeca84e795e754cd28a2", "output_hash": "30b20d9aae4d7b0587edeefdc2a9a159f73faf6e2570f1d408a415fe92cbf45e"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "4f966ec893d730ab8393d2e6", "attempt_id": "34d5a45f5cc3", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:32:39.729827Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 30}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 1242, "message_tokens": 1149, "message_utf8_bytes": 3043, "cache_tensor_bytes": null, "serialized_payload_bytes": 3043, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27333.36341707036, "ttft_ms": 13.823208981193602, "decode_ms": 14.106916962191463, "receiver_prefill_ms": 151.01191599387676}, "total_wall_ms": 27514.223334146664, "ttft_ms": 13.823208981193602, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "2dd52fe72ca47aacd0e383e165bc8f70c1b82d5f1667eeca84e795e754cd28a2", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "981b0f71877380b1257db25c", "attempt_id": "27760b4ba384", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:32:39.925244Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 30}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 1249, "message_tokens": 1149, "message_utf8_bytes": 3043, "cache_tensor_bytes": null, "serialized_payload_bytes": 3043, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27333.36341707036, "ttft_ms": 13.963583041913807, "decode_ms": 124.56249992828816, "receiver_prefill_ms": 193.1782920146361}, "total_wall_ms": 27667.046417016536, "ttft_ms": 13.963583041913807, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "6,786,800", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "2dd52fe72ca47aacd0e383e165bc8f70c1b82d5f1667eeca84e795e754cd28a2", "output_hash": "396a4b6b618fb10a20ca84793c54e794cbfebea7547dd351971f763f30153ef9"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b27b935d778dc2335edab948", "attempt_id": "105f8bef7472", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:32:40.274052Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 30}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 1240, "message_tokens": 1149, "message_utf8_bytes": 3043, "cache_tensor_bytes": null, "serialized_payload_bytes": 3043, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27333.36341707036, "ttft_ms": 13.867667061276734, "decode_ms": 124.35566692147404, "receiver_prefill_ms": 128.4382080193609}, "total_wall_ms": 27602.0553749986, "ttft_ms": 13.867667061276734, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8,810,500", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "2dd52fe72ca47aacd0e383e165bc8f70c1b82d5f1667eeca84e795e754cd28a2", "output_hash": "7810d312d2690f45bb13651b3c1cf9a2098b9e9d043463aaccc81157fad4a26e"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "5acab6d391204f7ee37aeac0", "attempt_id": "66add46b2d5b", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:32:40.558043Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 30}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 1250, "message_tokens": 1149, "message_utf8_bytes": 3043, "cache_tensor_bytes": null, "serialized_payload_bytes": 3043, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27333.36341707036, "ttft_ms": 13.429707963950932, "decode_ms": 80.71679202839732, "receiver_prefill_ms": 131.20533304754645}, "total_wall_ms": 27560.630958992988, "ttft_ms": 13.429707963950932, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Quidra-11", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "2dd52fe72ca47aacd0e383e165bc8f70c1b82d5f1667eeca84e795e754cd28a2", "output_hash": "41698c622e83fd669a8d097b3f00cb4197b0a66bac5aeca05b6224ada8b9dd5d"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "e89174e6870d3d56a576221e", "attempt_id": "58ef165faa03", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:32:40.799961Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 30}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 1243, "message_tokens": 1149, "message_utf8_bytes": 3043, "cache_tensor_bytes": null, "serialized_payload_bytes": 3043, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27333.36341707036, "ttft_ms": 13.370458036661148, "decode_ms": 80.1679160213098, "receiver_prefill_ms": 128.00354196224362}, "total_wall_ms": 27556.768625159748, "ttft_ms": 13.370458036661148, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "34,456", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "2dd52fe72ca47aacd0e383e165bc8f70c1b82d5f1667eeca84e795e754cd28a2", "output_hash": "ef314acff148203832c5f9820b067bf033e6f1e9d61af2610137dcdc1928db40"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "837d5ef543bdc6ca6cbc0442", "attempt_id": "6c830d487917", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:32:41.037897Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 30}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 1244, "message_tokens": 1149, "message_utf8_bytes": 3043, "cache_tensor_bytes": null, "serialized_payload_bytes": 3043, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27333.36341707036, "ttft_ms": 13.505125069059432, "decode_ms": 13.253499986603856, "receiver_prefill_ms": 128.2636660616845}, "total_wall_ms": 27490.23625010159, "ttft_ms": 13.505125069059432, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "2dd52fe72ca47aacd0e383e165bc8f70c1b82d5f1667eeca84e795e754cd28a2", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "9ed54487c05fdf7bfe473ef8", "attempt_id": "57b924af334b", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:32:41.209438Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 30}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 1245, "message_tokens": 1149, "message_utf8_bytes": 3043, "cache_tensor_bytes": null, "serialized_payload_bytes": 3043, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27333.36341707036, "ttft_ms": 13.831417076289654, "decode_ms": 13.499249937012792, "receiver_prefill_ms": 157.6941660605371}, "total_wall_ms": 27520.314209046774, "ttft_ms": 13.831417076289654, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "2dd52fe72ca47aacd0e383e165bc8f70c1b82d5f1667eeca84e795e754cd28a2", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "115139347f8d40f7670a8074", "attempt_id": "3fe27ffecb80", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:33:08.802256Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 2759, "message_tokens": 2671, "message_utf8_bytes": 6724, "cache_tensor_bytes": null, "serialized_payload_bytes": 6724, "side_information_bytes": 0, "fetched_bytes": 3489, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27390.31591603998, "ttft_ms": 13.56162503361702, "decode_ms": 93.48166605923325, "receiver_prefill_ms": 364.1218749107793}, "total_wall_ms": 27865.72549911216, "ttft_ms": 13.56162503361702, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "6009480", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "b5c328c6703a9adfbc34cc218c493f431c3c478abcb603a5ea119677417ddf88", "output_hash": "925ebdb42a0311d11ea73517fd780f133bed1e6fad2f9f6d66408fe5dba57a45"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "fb45f7affcd9a63d30a0cac2", "attempt_id": "0b7e587c24cc", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:33:09.293591Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 2644, "message_tokens": 2552, "message_utf8_bytes": 6521, "cache_tensor_bytes": null, "serialized_payload_bytes": 6521, "side_information_bytes": 0, "fetched_bytes": 3301, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27390.31591603998, "ttft_ms": 13.389874948188663, "decode_ms": 13.765415991656482, "receiver_prefill_ms": 615.7390000298619}, "total_wall_ms": 28037.718499079347, "ttft_ms": 13.389874948188663, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "cb74f20e80d0aa082fc4ef07751031b8a87cc2aacc58ba4250b1f5e89f0f0f74", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "8fa6be2d76944e07fe22b9a3", "attempt_id": "004ee6694d5f", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:33:09.955048Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 2775, "message_tokens": 2676, "message_utf8_bytes": 6703, "cache_tensor_bytes": null, "serialized_payload_bytes": 6703, "side_information_bytes": 0, "fetched_bytes": 3462, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27390.31591603998, "ttft_ms": 13.666291953995824, "decode_ms": 94.9016670929268, "receiver_prefill_ms": 300.5081668961793}, "total_wall_ms": 27802.976541104726, "ttft_ms": 13.666291953995824, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "6786800", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "d28ade61eca68ee15ba0bcf6521113abbb0eefa8a04b565841886f68071a5ce3", "output_hash": "33cee35795111ffd7720bcc954c699ed0ef5e51fe4ea1091cab2d69cd6de104f"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "8e83ba299111c312d5348105", "attempt_id": "15600aff359c", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:33:10.382301Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 2768, "message_tokens": 2678, "message_utf8_bytes": 6726, "cache_tensor_bytes": null, "serialized_payload_bytes": 6726, "side_information_bytes": 0, "fetched_bytes": 3489, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27390.31591603998, "ttft_ms": 13.705416000448167, "decode_ms": 94.56875000614673, "receiver_prefill_ms": 307.434624992311}, "total_wall_ms": 27809.710790985264, "ttft_ms": 13.705416000448167, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7510570", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "e1bd2a56cdee5be7792048ab273af9d6b860cb45cf1eb874302c11802cfee1ed", "output_hash": "410d7d10eb5e18b187a54949b6517bffab264255a8ac37e64c7fa5b57e44d9ed"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "5a32c2fefad51191c26dde0d", "attempt_id": "f0319ef0bdbe", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:33:10.816421Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 2628, "message_tokens": 2528, "message_utf8_bytes": 6436, "cache_tensor_bytes": null, "serialized_payload_bytes": 6436, "side_information_bytes": 0, "fetched_bytes": 3234, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27390.31591603998, "ttft_ms": 13.97454203106463, "decode_ms": 81.86804200522602, "receiver_prefill_ms": 409.29108299314976}, "total_wall_ms": 27901.343708042987, "ttft_ms": 13.97454203106463, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Quidra-11", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "4664b719443ee0ecc861740dd7a05468b98d7932904c796b36181a3681a1a024", "output_hash": "41698c622e83fd669a8d097b3f00cb4197b0a66bac5aeca05b6224ada8b9dd5d"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "ebef38725b83f9b48f2efb56", "attempt_id": "74803a156787", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:33:11.342875Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 2771, "message_tokens": 2678, "message_utf8_bytes": 6726, "cache_tensor_bytes": null, "serialized_payload_bytes": 6726, "side_information_bytes": 0, "fetched_bytes": 3489, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27390.31591603998, "ttft_ms": 14.187792083248496, "decode_ms": 109.44312496576458, "receiver_prefill_ms": 322.9266250273213}, "total_wall_ms": 27840.555041097105, "ttft_ms": 14.187792083248496, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "20874476", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "6152b64b5f4890cc599dcfb0232a7956fb48a2664ac40285a52f8f28a99fe87a", "output_hash": "3c88c562589ff9497bc35cfa0f4576277d138f78b6a52230ba2be63f01495c47"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "61fc895a2b27f30d747e7dcc", "attempt_id": "6f6e33322e21", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:33:11.808053Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 2661, "message_tokens": 2567, "message_utf8_bytes": 6590, "cache_tensor_bytes": null, "serialized_payload_bytes": 6590, "side_information_bytes": 0, "fetched_bytes": 3360, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27390.31591603998, "ttft_ms": 13.96333402954042, "decode_ms": 13.64699995610863, "receiver_prefill_ms": 319.23141691368073}, "total_wall_ms": 27740.73395796586, "ttft_ms": 13.96333402954042, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "05884cdf42dc217fcb72cba800a67716b54687b837beefbc489324a0dc36b2ff", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "e1c7d226d258eff644160b41", "attempt_id": "09a243683193", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:33:12.173429Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 2704, "message_tokens": 2609, "message_utf8_bytes": 6616, "cache_tensor_bytes": null, "serialized_payload_bytes": 6616, "side_information_bytes": 0, "fetched_bytes": 3394, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27390.31591603998, "ttft_ms": 13.551958021707833, "decode_ms": 13.408583006821573, "receiver_prefill_ms": 488.02433302626014}, "total_wall_ms": 27908.864041091874, "ttft_ms": 13.551958021707833, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "282f5bb1dd56a7033596c23a867f38ea470228b132503d88c14f31b081658549", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "bf154d799af69b14ec9eda56", "attempt_id": "8e16fa8f0e39", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:33:15.098925Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 98942976}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 42, "message_tokens": null, "message_utf8_bytes": 9014, "cache_tensor_bytes": 98942976, "serialized_payload_bytes": 98942976, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 2391.9723329599947, "ttft_ms": 31.792583991773427, "decode_ms": 129.5679589966312, "receiver_prefill_ms": 0.0}, "total_wall_ms": 2561.197373899631, "ttft_ms": 31.792583991773427, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8,810,500", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "7810d312d2690f45bb13651b3c1cf9a2098b9e9d043463aaccc81157fad4a26e"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "3a048188bf832953535db146", "attempt_id": "d0d12550bd12", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:33:15.282700Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 98942976}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 46, "message_tokens": null, "message_utf8_bytes": 9014, "cache_tensor_bytes": 98942976, "serialized_payload_bytes": 98942976, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 2391.9723329599947, "ttft_ms": 50.02279102336615, "decode_ms": 142.32412492856383, "receiver_prefill_ms": 0.0}, "total_wall_ms": 2587.387957959436, "ttft_ms": 50.02279102336615, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "CT-20T-7883", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "dd2e4bce724b0497351a41dbf71da774970bfcbe2553b73571a02bc7e8fa0e53"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b5a06d789f6c04b595055849", "attempt_id": "0dd07772880b", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:33:15.492688Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 98942976}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 53, "message_tokens": null, "message_utf8_bytes": 9014, "cache_tensor_bytes": 98942976, "serialized_payload_bytes": 98942976, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 2391.9723329599947, "ttft_ms": 45.61383300460875, "decode_ms": 99.45887490175664, "receiver_prefill_ms": 0.0}, "total_wall_ms": 2540.0656249839813, "ttft_ms": 45.61383300460875, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "6786800", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "33cee35795111ffd7720bcc954c699ed0ef5e51fe4ea1091cab2d69cd6de104f"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "d0a7c38db0e67b39395a1e3c", "attempt_id": "7aa9e89f0419", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:33:15.655240Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 98942976}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 44, "message_tokens": null, "message_utf8_bytes": 9014, "cache_tensor_bytes": 98942976, "serialized_payload_bytes": 98942976, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 2391.9723329599947, "ttft_ms": 55.29454105999321, "decode_ms": 127.93062499258667, "receiver_prefill_ms": 0.0}, "total_wall_ms": 2577.6857079472393, "ttft_ms": 55.29454105999321, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5,453,600", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "e6a2b25d3f99c6c8c339dc11738104244c59975cc325ebc53bc602e40b327216"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "f808aacd731bcd786576c46e", "attempt_id": "e3897edef821", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:33:15.855136Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 98942976}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 54, "message_tokens": null, "message_utf8_bytes": 9014, "cache_tensor_bytes": 98942976, "serialized_payload_bytes": 98942976, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 2391.9723329599947, "ttft_ms": 30.208124895580113, "decode_ms": 84.92729195859283, "receiver_prefill_ms": 0.0}, "total_wall_ms": 2510.1319579407573, "ttft_ms": 30.208124895580113, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Draard-12", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "c98ed1d4a10f5c2d70e15663c4ed0df74b420086785b671165017dfed27dee41"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a1bbc48f14740d0e4602d089", "attempt_id": "1afa8f9d0b43", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:33:15.987843Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 98942976}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 9014, "cache_tensor_bytes": 98942976, "serialized_payload_bytes": 98942976, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 2391.9723329599947, "ttft_ms": 30.716209090314806, "decode_ms": 142.4671249696985, "receiver_prefill_ms": 0.0}, "total_wall_ms": 2568.0527499644086, "ttft_ms": 30.716209090314806, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "14,342,972", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "0fc7a451f2136364ca437a706951fe46718490fee7c0150579caa1c401451cc5"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "eca92e8010179b22bb261707", "attempt_id": "4e62717a3d94", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:33:16.178020Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 98942976}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 48, "message_tokens": null, "message_utf8_bytes": 9014, "cache_tensor_bytes": 98942976, "serialized_payload_bytes": 98942976, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 2391.9723329599947, "ttft_ms": 69.29412495810539, "decode_ms": 14.1245829872787, "receiver_prefill_ms": 0.0}, "total_wall_ms": 2477.791832992807, "ttft_ms": 69.29412495810539, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "e4e5f0117f7980077ab3b0b8", "attempt_id": "b5c83a4f37b6", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:33:16.278136Z", "task_family": "synthetic", "document_id": "syn_03024_2000", "document_hash": "ea10cd11ad08f01c50b0712bc481668857c9d2ddc8e6423968f2a31eeb97ac55", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03024_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 98942976}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2068, "query_tokens": 49, "message_tokens": null, "message_utf8_bytes": 9014, "cache_tensor_bytes": 98942976, "serialized_payload_bytes": 98942976, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 2391.9723329599947, "ttft_ms": 30.782249988988042, "decode_ms": 14.210249995812774, "receiver_prefill_ms": 0.0}, "total_wall_ms": 2439.6806659642607, "ttft_ms": 30.782249988988042, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Quilun-16", "bogus_amount": "8,495,020", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03024_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "195c757c397eb4a92c81b463", "attempt_id": "cb32a702154e", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:33:17.338638Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 6282, "message_tokens": 6197, "message_utf8_bytes": 30751, "cache_tensor_bytes": null, "serialized_payload_bytes": 30751, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.03529200330376625, "ttft_ms": 14.566833968274295, "decode_ms": 124.82466700021178, "receiver_prefill_ms": 773.3494589338079}, "total_wall_ms": 924.3615000741556, "ttft_ms": 14.566833968274295, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5,236,625", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "6e71e6c5365db298b9144d66f0d893098a5ed3ab97d88cee8bbc2b5fd46de8fa", "output_hash": "5f724075674359492cf09f6cc38fa29d1d8d57b1cbe36c12f18a44145249093b"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "f86e069fdcb8f94789513b90", "attempt_id": "7940a44af8f2", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:33:18.279434Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 6285, "message_tokens": 6197, "message_utf8_bytes": 30751, "cache_tensor_bytes": null, "serialized_payload_bytes": 30751, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.03529200330376625, "ttft_ms": 13.964917045086622, "decode_ms": 137.56558403838426, "receiver_prefill_ms": 770.1070840703323}, "total_wall_ms": 931.5631260396913, "ttft_ms": 13.964917045086622, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "CN-53T-4564", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "6e71e6c5365db298b9144d66f0d893098a5ed3ab97d88cee8bbc2b5fd46de8fa", "output_hash": "63b72c2f0fa9c87c99863654eaed185c2db37aca56aeb2c04043723bedec7f95"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c00c55128d53f5ec6cc75bb0", "attempt_id": "869b2ab8dbf4", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:33:19.227766Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 6292, "message_tokens": 6197, "message_utf8_bytes": 30751, "cache_tensor_bytes": null, "serialized_payload_bytes": 30751, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.03529200330376625, "ttft_ms": 13.927624910138547, "decode_ms": 96.57708392478526, "receiver_prefill_ms": 787.4500410398468}, "total_wall_ms": 907.6572089688852, "ttft_ms": 13.927624910138547, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7026040", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "6e71e6c5365db298b9144d66f0d893098a5ed3ab97d88cee8bbc2b5fd46de8fa", "output_hash": "745005bc5caf64de51f389cd215c24264b3c2b8761c288554f09868d28ea3ca3"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "11f90642e22088d4e9a37522", "attempt_id": "6448c944137b", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:33:20.151806Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 6283, "message_tokens": 6197, "message_utf8_bytes": 30751, "cache_tensor_bytes": null, "serialized_payload_bytes": 30751, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.03529200330376625, "ttft_ms": 13.976292102597654, "decode_ms": 123.87224996928126, "receiver_prefill_ms": 785.5084170587361}, "total_wall_ms": 932.863583904691, "ttft_ms": 13.976292102597654, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4,824,670", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "6e71e6c5365db298b9144d66f0d893098a5ed3ab97d88cee8bbc2b5fd46de8fa", "output_hash": "39851ef0b05834836a8abff080829144cd3d975d77f3c92cc6a0a87722460d9a"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "4ddf773bd8fadef0d11b185d", "attempt_id": "72ba4cc6c311", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:33:21.101508Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 6293, "message_tokens": 6197, "message_utf8_bytes": 30751, "cache_tensor_bytes": null, "serialized_payload_bytes": 30751, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.03529200330376625, "ttft_ms": 14.33687494136393, "decode_ms": 69.58745792508125, "receiver_prefill_ms": 782.0380410412326}, "total_wall_ms": 875.6143749924377, "ttft_ms": 14.33687494136393, "input_usage": {}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Urker-16", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "6e71e6c5365db298b9144d66f0d893098a5ed3ab97d88cee8bbc2b5fd46de8fa", "output_hash": "13bcd50d62725e833c519fdb3eb450946e33fce2aca726f9c0f57cbe4cbcb2d9"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "7585759d1f2a8211def0e418", "attempt_id": "4a99b1c428e9", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:33:21.993843Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 6286, "message_tokens": 6197, "message_utf8_bytes": 30751, "cache_tensor_bytes": null, "serialized_payload_bytes": 30751, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.03529200330376625, "ttft_ms": 13.875333010219038, "decode_ms": 138.56929203029722, "receiver_prefill_ms": 790.5014170100912}, "total_wall_ms": 952.6870419504121, "ttft_ms": 13.875333010219038, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "20,882,147", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "6e71e6c5365db298b9144d66f0d893098a5ed3ab97d88cee8bbc2b5fd46de8fa", "output_hash": "356b17197275dcee7c65db2fe000a00f20e9b375cc19defdddb60dd096101eab"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "8b49097029d2688b9107a86e", "attempt_id": "db5d9509c519", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:33:22.963793Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 6286, "message_tokens": 6197, "message_utf8_bytes": 30751, "cache_tensor_bytes": null, "serialized_payload_bytes": 30751, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.03529200330376625, "ttft_ms": 13.919458026066422, "decode_ms": 14.11487499717623, "receiver_prefill_ms": 781.3797920243815}, "total_wall_ms": 819.1698759328574, "ttft_ms": 13.919458026066422, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "6e71e6c5365db298b9144d66f0d893098a5ed3ab97d88cee8bbc2b5fd46de8fa", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "eff3b72bdc0eaa80b6c3d221", "attempt_id": "76d3ccd11381", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:33:23.800026Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 6288, "message_tokens": 6197, "message_utf8_bytes": 30751, "cache_tensor_bytes": null, "serialized_payload_bytes": 30751, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.03529200330376625, "ttft_ms": 14.074750011786819, "decode_ms": 14.070541015826166, "receiver_prefill_ms": 805.3054999327287}, "total_wall_ms": 843.2813750114292, "ttft_ms": 14.074750011786819, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "6e71e6c5365db298b9144d66f0d893098a5ed3ab97d88cee8bbc2b5fd46de8fa", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "e482fe1bf4cbc40dd6c64c49", "attempt_id": "f86129725bf7", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:33:39.306993Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 1122, "message_tokens": 1036, "message_utf8_bytes": 2652, "cache_tensor_bytes": null, "serialized_payload_bytes": 2652, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14646.676875068806, "ttft_ms": 13.619957957416773, "decode_ms": 94.66941701248288, "receiver_prefill_ms": 121.49412499275059}, "total_wall_ms": 14883.74925008975, "ttft_ms": 13.619957957416773, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5236625", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "4b09a0c144966bbad4320908f48f320f64d6dff1f570edcaf7f18dbfc5132a04", "output_hash": "f0324cfa0e35a8d9bd4a75ef3f8cf62fcab822341c95b1d00f17b53fdf75ccf5"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "3ecf4921bca527f4b3b1610c", "attempt_id": "55a9068115b2", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:33:39.561858Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 1125, "message_tokens": 1036, "message_utf8_bytes": 2652, "cache_tensor_bytes": null, "serialized_payload_bytes": 2652, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14646.676875068806, "ttft_ms": 13.44216603320092, "decode_ms": 135.84958296269178, "receiver_prefill_ms": 121.03949999436736}, "total_wall_ms": 14918.862166116014, "ttft_ms": 13.44216603320092, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "CN-53T-4564", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "4b09a0c144966bbad4320908f48f320f64d6dff1f570edcaf7f18dbfc5132a04", "output_hash": "63b72c2f0fa9c87c99863654eaed185c2db37aca56aeb2c04043723bedec7f95"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "0b0d278957464f0c99bfe905", "attempt_id": "1e6dc0b3b2ec", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:33:39.851693Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 1132, "message_tokens": 1036, "message_utf8_bytes": 2652, "cache_tensor_bytes": null, "serialized_payload_bytes": 2652, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14646.676875068806, "ttft_ms": 13.620665995404124, "decode_ms": 95.00462491996586, "receiver_prefill_ms": 121.97562505025417}, "total_wall_ms": 14879.294709069654, "ttft_ms": 13.620665995404124, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7026040", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "4b09a0c144966bbad4320908f48f320f64d6dff1f570edcaf7f18dbfc5132a04", "output_hash": "745005bc5caf64de51f389cd215c24264b3c2b8761c288554f09868d28ea3ca3"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "25f7cf62018972da49464b8e", "attempt_id": "28ff1200b9a5", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:33:40.101676Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 1123, "message_tokens": 1036, "message_utf8_bytes": 2652, "cache_tensor_bytes": null, "serialized_payload_bytes": 2652, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14646.676875068806, "ttft_ms": 13.342624995857477, "decode_ms": 120.33591605722904, "receiver_prefill_ms": 121.00195896346122}, "total_wall_ms": 14903.166333097033, "ttft_ms": 13.342624995857477, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5,187,820", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "4b09a0c144966bbad4320908f48f320f64d6dff1f570edcaf7f18dbfc5132a04", "output_hash": "856804d5af6bf81ecc932fd68c47bd059046f4573ca64e21cb0c4c5d96dc59ca"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a16a256c3e46dca9e9d8dad5", "attempt_id": "40786d48c614", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:33:40.375152Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 1133, "message_tokens": 1036, "message_utf8_bytes": 2652, "cache_tensor_bytes": null, "serialized_payload_bytes": 2652, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14646.676875068806, "ttft_ms": 13.853291980922222, "decode_ms": 66.88429194036871, "receiver_prefill_ms": 121.03204207960516}, "total_wall_ms": 14850.352042121813, "ttft_ms": 13.853291980922222, "input_usage": {}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Urker-16", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "4b09a0c144966bbad4320908f48f320f64d6dff1f570edcaf7f18dbfc5132a04", "output_hash": "13bcd50d62725e833c519fdb3eb450946e33fce2aca726f9c0f57cbe4cbcb2d9"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "e7c4a0cbc034e0a16721896f", "attempt_id": "01a741d77407", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:33:40.596002Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 1126, "message_tokens": 1036, "message_utf8_bytes": 2652, "cache_tensor_bytes": null, "serialized_payload_bytes": 2652, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14646.676875068806, "ttft_ms": 13.264250010251999, "decode_ms": 134.02175006922334, "receiver_prefill_ms": 121.02229101583362}, "total_wall_ms": 14916.84162511956, "ttft_ms": 13.264250010251999, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "21,265,704", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "4b09a0c144966bbad4320908f48f320f64d6dff1f570edcaf7f18dbfc5132a04", "output_hash": "328730e89365ae840589e47ac26812af50340f1195419cdbf491ccd587a2eab5"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "f10f8be51b4d2045653abc65", "attempt_id": "68368e08792e", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:33:40.883189Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 1126, "message_tokens": 1036, "message_utf8_bytes": 2652, "cache_tensor_bytes": null, "serialized_payload_bytes": 2652, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14646.676875068806, "ttft_ms": 13.58512497972697, "decode_ms": 13.329374953173101, "receiver_prefill_ms": 121.03512499015778}, "total_wall_ms": 14796.446541091427, "ttft_ms": 13.58512497972697, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "4b09a0c144966bbad4320908f48f320f64d6dff1f570edcaf7f18dbfc5132a04", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b6741e964b29e6caddf62339", "attempt_id": "aa569d13045e", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:33:41.049743Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 1128, "message_tokens": 1036, "message_utf8_bytes": 2652, "cache_tensor_bytes": null, "serialized_payload_bytes": 2652, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14646.676875068806, "ttft_ms": 13.491875026375055, "decode_ms": 13.383207959122956, "receiver_prefill_ms": 121.07612495310605}, "total_wall_ms": 14796.496208989993, "ttft_ms": 13.491875026375055, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "4b09a0c144966bbad4320908f48f320f64d6dff1f570edcaf7f18dbfc5132a04", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "3a559c2bbf95cb30ade97a6d", "attempt_id": "65f989e6a54d", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:33:55.871482Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 17}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 766, "message_tokens": 676, "message_utf8_bytes": 1801, "cache_tensor_bytes": null, "serialized_payload_bytes": 1801, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14654.734582989477, "ttft_ms": 14.097583014518023, "decode_ms": 121.62129196804017, "receiver_prefill_ms": 81.88412489835173}, "total_wall_ms": 14874.55825007055, "ttft_ms": 14.097583014518023, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5,236,625", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "9544785caee37c111d4ccdd3e42c0f29fab5a7e269ece52dd01f8ed68f8e4b43", "output_hash": "5f724075674359492cf09f6cc38fa29d1d8d57b1cbe36c12f18a44145249093b"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "fff3de4201f2abd447d105fe", "attempt_id": "fc36aef04381", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:33:56.112012Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 17}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 769, "message_tokens": 676, "message_utf8_bytes": 1801, "cache_tensor_bytes": null, "serialized_payload_bytes": 1801, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14654.734582989477, "ttft_ms": 13.71258299332112, "decode_ms": 13.39779095724225, "receiver_prefill_ms": 78.58491700608283}, "total_wall_ms": 14761.902332887985, "ttft_ms": 13.71258299332112, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "9544785caee37c111d4ccdd3e42c0f29fab5a7e269ece52dd01f8ed68f8e4b43", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "681caab65502ccd12dc3ff21", "attempt_id": "487a6810e588", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:33:56.236176Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 17}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 776, "message_tokens": 676, "message_utf8_bytes": 1801, "cache_tensor_bytes": null, "serialized_payload_bytes": 1801, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14654.734582989477, "ttft_ms": 13.510874938219786, "decode_ms": 93.41791598126292, "receiver_prefill_ms": 92.39491599146277}, "total_wall_ms": 14855.440167011693, "ttft_ms": 13.510874938219786, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8252854", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "9544785caee37c111d4ccdd3e42c0f29fab5a7e269ece52dd01f8ed68f8e4b43", "output_hash": "0b172441eb5f315b0e90c34cabeb1d228dce23919fa88a7bb2734d342ff1b48e"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "1c50f43976ef1d785a5cdabd", "attempt_id": "ab7e8d708055", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:33:56.453798Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 17}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 767, "message_tokens": 676, "message_utf8_bytes": 1801, "cache_tensor_bytes": null, "serialized_payload_bytes": 1801, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14654.734582989477, "ttft_ms": 13.295584009028971, "decode_ms": 94.72499997355044, "receiver_prefill_ms": 81.62887499202043}, "total_wall_ms": 14845.768040977418, "ttft_ms": 13.295584009028971, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "6929235", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "9544785caee37c111d4ccdd3e42c0f29fab5a7e269ece52dd01f8ed68f8e4b43", "output_hash": "aae99ad47be986117fc352fa40a60c128f3f628966ca197ee1b83d1081f56970"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c341a6c49fdb21edb2dfafbd", "attempt_id": "e1ad0ce1f70b", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:33:56.662375Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 17}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 777, "message_tokens": 676, "message_utf8_bytes": 1801, "cache_tensor_bytes": null, "serialized_payload_bytes": 1801, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14654.734582989477, "ttft_ms": 13.195708044804633, "decode_ms": 66.65245804470032, "receiver_prefill_ms": 91.83295804541558}, "total_wall_ms": 14827.782499953173, "ttft_ms": 13.195708044804633, "input_usage": {}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Urker-16", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "9544785caee37c111d4ccdd3e42c0f29fab5a7e269ece52dd01f8ed68f8e4b43", "output_hash": "13bcd50d62725e833c519fdb3eb450946e33fce2aca726f9c0f57cbe4cbcb2d9"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "8fcdd7cafe6ab37ff1d40721", "attempt_id": "e6204e34b8a2", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:33:56.852550Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 17}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 770, "message_tokens": 676, "message_utf8_bytes": 1801, "cache_tensor_bytes": null, "serialized_payload_bytes": 1801, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14654.734582989477, "ttft_ms": 13.555125100538135, "decode_ms": 133.0276249209419, "receiver_prefill_ms": 90.84954194258898}, "total_wall_ms": 14893.59804103151, "ttft_ms": 13.555125100538135, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "21,078,704", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "9544785caee37c111d4ccdd3e42c0f29fab5a7e269ece52dd01f8ed68f8e4b43", "output_hash": "a2af1ef79d7b65c04f37afbc85d8da9676a07e54e60debe5305645ee5912191f"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "df1d357447ed5e5f7909842d", "attempt_id": "43bf75f9f537", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:33:57.108691Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 17}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 770, "message_tokens": 676, "message_utf8_bytes": 1801, "cache_tensor_bytes": null, "serialized_payload_bytes": 1801, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14654.734582989477, "ttft_ms": 13.2884579943493, "decode_ms": 13.697791961021721, "receiver_prefill_ms": 90.24091600440443}, "total_wall_ms": 14773.334875004366, "ttft_ms": 13.2884579943493, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "9544785caee37c111d4ccdd3e42c0f29fab5a7e269ece52dd01f8ed68f8e4b43", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "acf073998ae9ee83c292f681", "attempt_id": "7a709b08748f", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:33:57.244551Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 17}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 772, "message_tokens": 676, "message_utf8_bytes": 1801, "cache_tensor_bytes": null, "serialized_payload_bytes": 1801, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14654.734582989477, "ttft_ms": 13.44474998768419, "decode_ms": 13.141416013240814, "receiver_prefill_ms": 90.89054109063}, "total_wall_ms": 14773.615416022949, "ttft_ms": 13.44474998768419, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "9544785caee37c111d4ccdd3e42c0f29fab5a7e269ece52dd01f8ed68f8e4b43", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "dada5c8828b37ddb203c3362", "attempt_id": "53775dbabe7e", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:34:12.036349Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 790, "message_tokens": 701, "message_utf8_bytes": 1757, "cache_tensor_bytes": null, "serialized_payload_bytes": 1757, "side_information_bytes": 0, "fetched_bytes": 929, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14655.413042055443, "ttft_ms": 13.614041963592172, "decode_ms": 93.29129196703434, "receiver_prefill_ms": 92.18445792794228}, "total_wall_ms": 14858.881958993152, "ttft_ms": 13.614041963592172, "input_usage": {"lookups": 3}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5236625", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "366cc92003006099d1b8b9223281ca2db73f81ff5b4f6c7c6b8fd927e3c18531", "output_hash": "f0324cfa0e35a8d9bd4a75ef3f8cf62fcab822341c95b1d00f17b53fdf75ccf5"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "60196e797f5ad41bc14f2a7f", "attempt_id": "92aaef7e2e5d", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:34:12.257624Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 738, "message_tokens": 646, "message_utf8_bytes": 1637, "cache_tensor_bytes": null, "serialized_payload_bytes": 1637, "side_information_bytes": 0, "fetched_bytes": 826, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14655.413042055443, "ttft_ms": 13.218333013355732, "decode_ms": 13.145708944648504, "receiver_prefill_ms": 81.08454197645187}, "total_wall_ms": 14764.273501001298, "ttft_ms": 13.218333013355732, "input_usage": {"lookups": 3}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "6dab76262d47b0e892084d7932fee9ebc6710b1693be25d2d5a7fcedc7e408fd", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c6daa96c3cdcc49fafd4442d", "attempt_id": "83eb6da68b0b", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:34:12.383318Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 787, "message_tokens": 688, "message_utf8_bytes": 1715, "cache_tensor_bytes": null, "serialized_payload_bytes": 1715, "side_information_bytes": 0, "fetched_bytes": 892, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14655.413042055443, "ttft_ms": 13.35245801601559, "decode_ms": 13.383499928750098, "receiver_prefill_ms": 91.95645805448294}, "total_wall_ms": 14775.595792103559, "ttft_ms": 13.35245801601559, "input_usage": {"lookups": 3}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "2a1bcd6dd7ebd79f8472f59bd051c9978b79f409cc83fb1dd32c35b96bda6e8c", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "f7ee7ffe21857216a14718e0", "attempt_id": "0d9716ba5e54", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:34:12.521124Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 736, "message_tokens": 646, "message_utf8_bytes": 1637, "cache_tensor_bytes": null, "serialized_payload_bytes": 1637, "side_information_bytes": 0, "fetched_bytes": 826, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14655.413042055443, "ttft_ms": 13.387083075940609, "decode_ms": 92.91187499184161, "receiver_prefill_ms": 79.47687502019107}, "total_wall_ms": 14842.623875010759, "ttft_ms": 13.387083075940609, "input_usage": {"lookups": 3}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5187820", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "36e63f5c10e1cc7ea5796e91edf32a5c0e82fc13d3e5ad6d1ff7ae6be98361e7", "output_hash": "506e6115b34b564401e129c2ef42e2e7e0ec82b154b45d92c85846383969e6ed"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "ac6cab49103603be11f8c72b", "attempt_id": "f84ec5795a4a", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:34:12.725940Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 773, "message_tokens": 673, "message_utf8_bytes": 1717, "cache_tensor_bytes": null, "serialized_payload_bytes": 1717, "side_information_bytes": 0, "fetched_bytes": 890, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14655.413042055443, "ttft_ms": 13.57708394061774, "decode_ms": 93.84529199451208, "receiver_prefill_ms": 90.84679197985679}, "total_wall_ms": 14855.115667101927, "ttft_ms": 13.57708394061774, "input_usage": {"lookups": 3}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Wentisoby-18", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ef6dc0f093d324bd5fec2def5091816957bc10f559c0317364759dd7dae06b81", "output_hash": "5e1b4376cbc7848ad56e533260fa5c8b3f9bbb233da09024ec193283565e0538"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "19276c72af0b83c0a1b7feb2", "attempt_id": "09833a3d75ee", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:34:12.942593Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 794, "message_tokens": 701, "message_utf8_bytes": 1757, "cache_tensor_bytes": null, "serialized_payload_bytes": 1757, "side_information_bytes": 0, "fetched_bytes": 929, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14655.413042055443, "ttft_ms": 13.43862502835691, "decode_ms": 134.01937508024275, "receiver_prefill_ms": 91.67170792352408}, "total_wall_ms": 14896.150959073566, "ttft_ms": 13.43862502835691, "input_usage": {"lookups": 3}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "14,017,885", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "366cc92003006099d1b8b9223281ca2db73f81ff5b4f6c7c6b8fd927e3c18531", "output_hash": "3aac6d7ef8ee4a890fd1ffbaeec7b2cc491e6791e284a6d1549cc189b5d6d07e"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "d79559c0bbb7b55f88834349", "attempt_id": "aaa1c55c582c", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:34:13.200621Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 781, "message_tokens": 688, "message_utf8_bytes": 1715, "cache_tensor_bytes": null, "serialized_payload_bytes": 1715, "side_information_bytes": 0, "fetched_bytes": 892, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14655.413042055443, "ttft_ms": 13.283749925903976, "decode_ms": 13.18645803257823, "receiver_prefill_ms": 91.93645801860839}, "total_wall_ms": 14775.302042020485, "ttft_ms": 13.283749925903976, "input_usage": {"lookups": 3}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "2a1bcd6dd7ebd79f8472f59bd051c9978b79f409cc83fb1dd32c35b96bda6e8c", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "9fd80ed77ae99018aab204b7", "attempt_id": "2b1b4e57d69e", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:34:13.337997Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 783, "message_tokens": 688, "message_utf8_bytes": 1715, "cache_tensor_bytes": null, "serialized_payload_bytes": 1715, "side_information_bytes": 0, "fetched_bytes": 892, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14655.413042055443, "ttft_ms": 14.13133298046887, "decode_ms": 13.12083296943456, "receiver_prefill_ms": 92.93558297213167}, "total_wall_ms": 14777.036417042837, "ttft_ms": 14.13133298046887, "input_usage": {"lookups": 3}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "2a1bcd6dd7ebd79f8472f59bd051c9978b79f409cc83fb1dd32c35b96bda6e8c", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "4207e74a2f4b3ebb21063735", "attempt_id": "213f350615e4", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:34:14.196230Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 231321600}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 43, "message_tokens": null, "message_utf8_bytes": 30725, "cache_tensor_bytes": 231321600, "serialized_payload_bytes": 231321600, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 719.3774590268731, "ttft_ms": 53.951416979543865, "decode_ms": 106.7062079673633, "receiver_prefill_ms": 0.0}, "total_wall_ms": 889.353625010699, "ttft_ms": 53.951416979543865, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5236625", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "f0324cfa0e35a8d9bd4a75ef3f8cf62fcab822341c95b1d00f17b53fdf75ccf5"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "e9032ad0d59f290c37ee7add", "attempt_id": "2031a02701ca", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:34:14.383238Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 231321600}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 46, "message_tokens": null, "message_utf8_bytes": 30725, "cache_tensor_bytes": 231321600, "serialized_payload_bytes": 231321600, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 719.3774590268731, "ttft_ms": 49.42712502088398, "decode_ms": 149.9504999956116, "receiver_prefill_ms": 0.0}, "total_wall_ms": 926.240042084828, "ttft_ms": 49.42712502088398, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "CN-53T-4564", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "63b72c2f0fa9c87c99863654eaed185c2db37aca56aeb2c04043723bedec7f95"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "f44c5e81d6a04f7f967c5911", "attempt_id": "ad3ff8d3ca52", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:34:14.608114Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 231321600}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 53, "message_tokens": null, "message_utf8_bytes": 30725, "cache_tensor_bytes": 231321600, "serialized_payload_bytes": 231321600, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 719.3774590268731, "ttft_ms": 46.42125009559095, "decode_ms": 104.94633298367262, "receiver_prefill_ms": 0.0}, "total_wall_ms": 878.5000429488719, "ttft_ms": 46.42125009559095, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7026040", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "745005bc5caf64de51f389cd215c24264b3c2b8761c288554f09868d28ea3ca3"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "e3f98f2118402001cb1f35e1", "attempt_id": "5b18d92496a1", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:34:14.784311Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 231321600}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 44, "message_tokens": null, "message_utf8_bytes": 30725, "cache_tensor_bytes": 231321600, "serialized_payload_bytes": 231321600, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 719.3774590268731, "ttft_ms": 47.992166015319526, "decode_ms": 105.60887504834682, "receiver_prefill_ms": 0.0}, "total_wall_ms": 880.6046260287985, "ttft_ms": 47.992166015319526, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5187820", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "506e6115b34b564401e129c2ef42e2e7e0ec82b154b45d92c85846383969e6ed"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "2c442f9c364fa0ed9de3af7a", "attempt_id": "f66c005a13b0", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:34:14.962224Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 231321600}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 54, "message_tokens": null, "message_utf8_bytes": 30725, "cache_tensor_bytes": 231321600, "serialized_payload_bytes": 231321600, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 719.3774590268731, "ttft_ms": 49.77920895908028, "decode_ms": 105.19916692283005, "receiver_prefill_ms": 0.0}, "total_wall_ms": 881.7845840239897, "ttft_ms": 49.77920895908028, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Wentisoby-18", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "5e1b4376cbc7848ad56e533260fa5c8b3f9bbb233da09024ec193283565e0538"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "5e3a52c900710b296e62168a", "attempt_id": "a07a778cd051", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:34:15.142200Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 231321600}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 30725, "cache_tensor_bytes": 231321600, "serialized_payload_bytes": 231321600, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 719.3774590268731, "ttft_ms": 47.02920909039676, "decode_ms": 120.07075001019984, "receiver_prefill_ms": 0.0}, "total_wall_ms": 894.1349590895697, "ttft_ms": 47.02920909039676, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "24084468", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "9cff8f9b32e3b9d064c4e2dfc1b66e78a731caaabf12e700bb3645c5725f9b6d"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "e57b3fe7d6eaa8e288ab6a58", "attempt_id": "982a20849ca1", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:34:15.334100Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 231321600}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 30725, "cache_tensor_bytes": 231321600, "serialized_payload_bytes": 231321600, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 719.3774590268731, "ttft_ms": 47.18604206573218, "decode_ms": 14.914250001311302, "receiver_prefill_ms": 0.0}, "total_wall_ms": 789.118001004681, "ttft_ms": 47.18604206573218, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b3112ff5e11a757bdf3c2104", "attempt_id": "143bb97d10b1", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:34:15.421171Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 231321600}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 49, "message_tokens": null, "message_utf8_bytes": 30725, "cache_tensor_bytes": 231321600, "serialized_payload_bytes": 231321600, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 719.3774590268731, "ttft_ms": 46.21583304833621, "decode_ms": 14.879208989441395, "receiver_prefill_ms": 0.0}, "total_wall_ms": 788.1164590362459, "ttft_ms": 46.21583304833621, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "9f4a53fb97b88f82b5bf4d90", "attempt_id": "22a5fdb7c664", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:34:15.507518Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 6396, "message_tokens": 6311, "message_utf8_bytes": 31255, "cache_tensor_bytes": null, "serialized_payload_bytes": 31255, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.026416033506393433, "ttft_ms": 13.842334039509296, "decode_ms": 98.42737508006394, "receiver_prefill_ms": 742.0234159799293}, "total_wall_ms": 867.1982489759102, "ttft_ms": 13.842334039509296, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5236625", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "deb638ddcef202b0a65018d3aa3f047f59ce22d9e9e2edab3549792e70b0d2e9", "output_hash": "f0324cfa0e35a8d9bd4a75ef3f8cf62fcab822341c95b1d00f17b53fdf75ccf5"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "01ac1ee08afe9a1a96c46b8f", "attempt_id": "42aae0ba707c", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:34:16.391273Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 6399, "message_tokens": 6311, "message_utf8_bytes": 31255, "cache_tensor_bytes": null, "serialized_payload_bytes": 31255, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.026416033506393433, "ttft_ms": 14.065707917325199, "decode_ms": 139.32808302342892, "receiver_prefill_ms": 749.9597080750391}, "total_wall_ms": 913.0894580157474, "ttft_ms": 14.065707917325199, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "CN-53T-4564", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "deb638ddcef202b0a65018d3aa3f047f59ce22d9e9e2edab3549792e70b0d2e9", "output_hash": "63b72c2f0fa9c87c99863654eaed185c2db37aca56aeb2c04043723bedec7f95"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "0500a1f56639e3d15becc936", "attempt_id": "9c2414822aa6", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:34:17.321687Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 6406, "message_tokens": 6311, "message_utf8_bytes": 31255, "cache_tensor_bytes": null, "serialized_payload_bytes": 31255, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.026416033506393433, "ttft_ms": 14.056374900974333, "decode_ms": 97.95545798260719, "receiver_prefill_ms": 773.3983750222251}, "total_wall_ms": 895.126624032855, "ttft_ms": 14.056374900974333, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7026040", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "deb638ddcef202b0a65018d3aa3f047f59ce22d9e9e2edab3549792e70b0d2e9", "output_hash": "745005bc5caf64de51f389cd215c24264b3c2b8761c288554f09868d28ea3ca3"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "881ebb5299f98bda55192b06", "attempt_id": "3310be822511", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:34:18.234005Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 6397, "message_tokens": 6311, "message_utf8_bytes": 31255, "cache_tensor_bytes": null, "serialized_payload_bytes": 31255, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.026416033506393433, "ttft_ms": 13.985792058520019, "decode_ms": 98.57400006148964, "receiver_prefill_ms": 770.3873330028728}, "total_wall_ms": 892.9402909707278, "ttft_ms": 13.985792058520019, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4824670", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "deb638ddcef202b0a65018d3aa3f047f59ce22d9e9e2edab3549792e70b0d2e9", "output_hash": "7d650cb178296568272fceaac0823be462a75f937d2aa9cd7e0fd6dc154adfb5"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "6f9bd5b720b908b211a17a2d", "attempt_id": "80644c6d0c0e", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:34:19.144328Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 6407, "message_tokens": 6311, "message_utf8_bytes": 31255, "cache_tensor_bytes": null, "serialized_payload_bytes": 31255, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.026416033506393433, "ttft_ms": 14.005624921992421, "decode_ms": 69.0582919633016, "receiver_prefill_ms": 791.1937079625204}, "total_wall_ms": 884.3467910774052, "ttft_ms": 14.005624921992421, "input_usage": {}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Urker-16", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "deb638ddcef202b0a65018d3aa3f047f59ce22d9e9e2edab3549792e70b0d2e9", "output_hash": "13bcd50d62725e833c519fdb3eb450946e33fce2aca726f9c0f57cbe4cbcb2d9"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "8bcceabb512dae91e978ee13", "attempt_id": "767e7c93d1bf", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:34:20.046088Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 6400, "message_tokens": 6311, "message_utf8_bytes": 31255, "cache_tensor_bytes": null, "serialized_payload_bytes": 31255, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.026416033506393433, "ttft_ms": 14.076124993152916, "decode_ms": 98.29187497962266, "receiver_prefill_ms": 783.9830829761922}, "total_wall_ms": 906.2285410473123, "ttft_ms": 14.076124993152916, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "2108194", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "deb638ddcef202b0a65018d3aa3f047f59ce22d9e9e2edab3549792e70b0d2e9", "output_hash": "b164dfbc1313f2b24b036772a55a1784d2c1cacd0f06555131578a7017216de8"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "5e40f7df7190d64b2dba8c3d", "attempt_id": "adbb00329794", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:34:20.969465Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 6400, "message_tokens": 6311, "message_utf8_bytes": 31255, "cache_tensor_bytes": null, "serialized_payload_bytes": 31255, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.026416033506393433, "ttft_ms": 14.023374998942018, "decode_ms": 15.310125076211989, "receiver_prefill_ms": 788.8275830773637}, "total_wall_ms": 828.0391240259632, "ttft_ms": 14.023374998942018, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "deb638ddcef202b0a65018d3aa3f047f59ce22d9e9e2edab3549792e70b0d2e9", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "ab6f4779a5f29442ae91005a", "attempt_id": "c4d7c322a3ef", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:34:21.814378Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 6402, "message_tokens": 6311, "message_utf8_bytes": 31255, "cache_tensor_bytes": null, "serialized_payload_bytes": 31255, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.026416033506393433, "ttft_ms": 13.9586670557037, "decode_ms": 13.951750006526709, "receiver_prefill_ms": 817.8779169684276}, "total_wall_ms": 855.4572910070419, "ttft_ms": 13.9586670557037, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "deb638ddcef202b0a65018d3aa3f047f59ce22d9e9e2edab3549792e70b0d2e9", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b4cd28e5ddaadd68119198c0", "attempt_id": "effe3d40daab", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:34:49.272276Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1918}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 2031, "message_tokens": 1945, "message_utf8_bytes": 7028, "cache_tensor_bytes": null, "serialized_payload_bytes": 7028, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26584.533707937226, "ttft_ms": 13.674458023160696, "decode_ms": 94.1409170627594, "receiver_prefill_ms": 211.16850001271814}, "total_wall_ms": 26912.351207924075, "ttft_ms": 13.674458023160696, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5236625", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a50f81a2f8d61b175a223107514bcd3e47489f9629aa63ee53e3ef597f48676c", "output_hash": "f0324cfa0e35a8d9bd4a75ef3f8cf62fcab822341c95b1d00f17b53fdf75ccf5"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "6c4267974cf670fb41f807f1", "attempt_id": "c081bc052ee3", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:34:49.620779Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1918}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 2034, "message_tokens": 1945, "message_utf8_bytes": 7028, "cache_tensor_bytes": null, "serialized_payload_bytes": 7028, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26584.533707937226, "ttft_ms": 13.631707988679409, "decode_ms": 136.76766608841717, "receiver_prefill_ms": 211.83491603005677}, "total_wall_ms": 26950.05374995526, "ttft_ms": 13.631707988679409, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "CN-53T-4564", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a50f81a2f8d61b175a223107514bcd3e47489f9629aa63ee53e3ef597f48676c", "output_hash": "63b72c2f0fa9c87c99863654eaed185c2db37aca56aeb2c04043723bedec7f95"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "5196fe27a2b1d0a015c22415", "attempt_id": "23de4c1bee7a", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:34:50.003571Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1918}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 2041, "message_tokens": 1945, "message_utf8_bytes": 7028, "cache_tensor_bytes": null, "serialized_payload_bytes": 7028, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26584.533707937226, "ttft_ms": 13.473458006046712, "decode_ms": 95.0988749973476, "receiver_prefill_ms": 212.26954204030335}, "total_wall_ms": 26908.625249983743, "ttft_ms": 13.473458006046712, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7026040", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a50f81a2f8d61b175a223107514bcd3e47489f9629aa63ee53e3ef597f48676c", "output_hash": "745005bc5caf64de51f389cd215c24264b3c2b8761c288554f09868d28ea3ca3"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "495a008cc883c5f3e9f5494c", "attempt_id": "73e8c75ce379", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:34:50.344976Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1918}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 2032, "message_tokens": 1945, "message_utf8_bytes": 7028, "cache_tensor_bytes": null, "serialized_payload_bytes": 7028, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26584.533707937226, "ttft_ms": 13.581625069491565, "decode_ms": 120.79645798075944, "receiver_prefill_ms": 211.27512503881007}, "total_wall_ms": 26933.40287497267, "ttft_ms": 13.581625069491565, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5,187,820", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a50f81a2f8d61b175a223107514bcd3e47489f9629aa63ee53e3ef597f48676c", "output_hash": "856804d5af6bf81ecc932fd68c47bd059046f4573ca64e21cb0c4c5d96dc59ca"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "302854827a7840cd87070863", "attempt_id": "855b371ddbbd", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:34:50.711600Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1918}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 2042, "message_tokens": 1945, "message_utf8_bytes": 7028, "cache_tensor_bytes": null, "serialized_payload_bytes": 7028, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26584.533707937226, "ttft_ms": 13.53633392136544, "decode_ms": 68.31995793618262, "receiver_prefill_ms": 212.65508304350078}, "total_wall_ms": 26882.301790989004, "ttft_ms": 13.53633392136544, "input_usage": {}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Urker-16", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a50f81a2f8d61b175a223107514bcd3e47489f9629aa63ee53e3ef597f48676c", "output_hash": "13bcd50d62725e833c519fdb3eb450946e33fce2aca726f9c0f57cbe4cbcb2d9"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a2d663921c6d7382e5ee8855", "attempt_id": "cb34f516bf5e", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:34:51.026590Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1918}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 2035, "message_tokens": 1945, "message_utf8_bytes": 7028, "cache_tensor_bytes": null, "serialized_payload_bytes": 7028, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26584.533707937226, "ttft_ms": 13.59320804476738, "decode_ms": 133.9391670189798, "receiver_prefill_ms": 211.96112490724772}, "total_wall_ms": 26947.249291930348, "ttft_ms": 13.59320804476738, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "20,203,342", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a50f81a2f8d61b175a223107514bcd3e47489f9629aa63ee53e3ef597f48676c", "output_hash": "a44be814c77bb38454e9e08c373e74291169b956d647a0230fe13778d105b7e5"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "4e798d7fd2551215a65e197e", "attempt_id": "4e20296e9a90", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:34:51.406706Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1918}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 2035, "message_tokens": 1945, "message_utf8_bytes": 7028, "cache_tensor_bytes": null, "serialized_payload_bytes": 7028, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26584.533707937226, "ttft_ms": 13.448459096252918, "decode_ms": 13.630874920636415, "receiver_prefill_ms": 211.191458045505}, "total_wall_ms": 26825.981457950547, "ttft_ms": 13.448459096252918, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a50f81a2f8d61b175a223107514bcd3e47489f9629aa63ee53e3ef597f48676c", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "464e43bbe8849f430d3581a0", "attempt_id": "d33d4014a28b", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:34:51.665639Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1918}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 2037, "message_tokens": 1945, "message_utf8_bytes": 7028, "cache_tensor_bytes": null, "serialized_payload_bytes": 7028, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 26584.533707937226, "ttft_ms": 13.886458007618785, "decode_ms": 13.766709016636014, "receiver_prefill_ms": 219.90108292084187}, "total_wall_ms": 26835.42541693896, "ttft_ms": 13.886458007618785, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a50f81a2f8d61b175a223107514bcd3e47489f9629aa63ee53e3ef597f48676c", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "45389cae005fb5b47bdf2a5d", "attempt_id": "3c58e708c1e6", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:35:20.145514Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 32}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 1351, "message_tokens": 1261, "message_utf8_bytes": 3395, "cache_tensor_bytes": null, "serialized_payload_bytes": 3395, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 28211.461542057805, "ttft_ms": 13.765792013145983, "decode_ms": 122.56724992766976, "receiver_prefill_ms": 154.54591694287956}, "total_wall_ms": 28505.102584022097, "ttft_ms": 13.765792013145983, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8,252,854", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a47d2d9e37e8b4359e1c296f914ab9bec64c3ebe6d0c0381d50b8e0b27f8f7e2", "output_hash": "9209da831fe61c3eb8d1a594ba822bc50ecd6afd78e837e992d96cc231d95e0c"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "774b610f675e7143c2331a85", "attempt_id": "adf71987f67c", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:35:20.457111Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 32}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 1354, "message_tokens": 1261, "message_utf8_bytes": 3395, "cache_tensor_bytes": null, "serialized_payload_bytes": 3395, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 28211.461542057805, "ttft_ms": 13.672124943695962, "decode_ms": 136.35337504092604, "receiver_prefill_ms": 144.8562090517953}, "total_wall_ms": 28508.454584050924, "ttft_ms": 13.672124943695962, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "CN-53T-4564", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a47d2d9e37e8b4359e1c296f914ab9bec64c3ebe6d0c0381d50b8e0b27f8f7e2", "output_hash": "63b72c2f0fa9c87c99863654eaed185c2db37aca56aeb2c04043723bedec7f95"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "2bc63e9268f46fec4257af5c", "attempt_id": "4b983c29d8ac", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:35:20.771658Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 32}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 1361, "message_tokens": 1261, "message_utf8_bytes": 3395, "cache_tensor_bytes": null, "serialized_payload_bytes": 3395, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 28211.461542057805, "ttft_ms": 14.183167018927634, "decode_ms": 97.51208301167935, "receiver_prefill_ms": 194.85087506473064}, "total_wall_ms": 28520.30579210259, "ttft_ms": 14.183167018927634, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8252854", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a47d2d9e37e8b4359e1c296f914ab9bec64c3ebe6d0c0381d50b8e0b27f8f7e2", "output_hash": "0b172441eb5f315b0e90c34cabeb1d228dce23919fa88a7bb2734d342ff1b48e"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "39031e0f4b5b9abcba3dffc3", "attempt_id": "0ffcf650293c", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:35:21.116233Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 32}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 1352, "message_tokens": 1261, "message_utf8_bytes": 3395, "cache_tensor_bytes": null, "serialized_payload_bytes": 3395, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 28211.461542057805, "ttft_ms": 13.656415976583958, "decode_ms": 96.27470897976309, "receiver_prefill_ms": 144.79699998628348}, "total_wall_ms": 28468.42879196629, "ttft_ms": 13.656415976583958, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "6069490", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a47d2d9e37e8b4359e1c296f914ab9bec64c3ebe6d0c0381d50b8e0b27f8f7e2", "output_hash": "4711441f1790872d27e09a73af4b4fe9b84256b0337928d055ea64ff622daa71"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c2c1109afbe4808b307c721e", "attempt_id": "ef3fd3665b5a", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:35:21.392273Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 32}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 1362, "message_tokens": 1261, "message_utf8_bytes": 3395, "cache_tensor_bytes": null, "serialized_payload_bytes": 3395, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 28211.461542057805, "ttft_ms": 13.831958058290184, "decode_ms": 69.74966591224074, "receiver_prefill_ms": 150.79145901836455}, "total_wall_ms": 28447.959584067576, "ttft_ms": 13.831958058290184, "input_usage": {}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Urker-16", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a47d2d9e37e8b4359e1c296f914ab9bec64c3ebe6d0c0381d50b8e0b27f8f7e2", "output_hash": "13bcd50d62725e833c519fdb3eb450946e33fce2aca726f9c0f57cbe4cbcb2d9"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "2599e29ca71bfaaa2f378337", "attempt_id": "f47a66099183", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:35:21.646798Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 32}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 1355, "message_tokens": 1261, "message_utf8_bytes": 3395, "cache_tensor_bytes": null, "serialized_payload_bytes": 3395, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 28211.461542057805, "ttft_ms": 13.70508293621242, "decode_ms": 136.20358298067003, "receiver_prefill_ms": 184.32254192885011}, "total_wall_ms": 28547.926958068274, "ttft_ms": 13.70508293621242, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "13,488,479", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a47d2d9e37e8b4359e1c296f914ab9bec64c3ebe6d0c0381d50b8e0b27f8f7e2", "output_hash": "6c6c60c4d336c7a8785c178f346d67b43a4c9b96a48c843bd7cd11784f64223e"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "14a3ad67ce4556d1cfb8c7b4", "attempt_id": "e73a495ceb08", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:35:22.001228Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 32}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 1355, "message_tokens": 1261, "message_utf8_bytes": 3395, "cache_tensor_bytes": null, "serialized_payload_bytes": 3395, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 28211.461542057805, "ttft_ms": 13.725124998018146, "decode_ms": 13.815375044941902, "receiver_prefill_ms": 144.33587493840605}, "total_wall_ms": 28385.604542098008, "ttft_ms": 13.725124998018146, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a47d2d9e37e8b4359e1c296f914ab9bec64c3ebe6d0c0381d50b8e0b27f8f7e2", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c50de12766ab77c341b156a9", "attempt_id": "31563b2fcbb6", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:35:22.193357Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 32}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 1357, "message_tokens": 1261, "message_utf8_bytes": 3395, "cache_tensor_bytes": null, "serialized_payload_bytes": 3395, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 28211.461542057805, "ttft_ms": 13.62950005568564, "decode_ms": 13.653457979671657, "receiver_prefill_ms": 177.26633290294558}, "total_wall_ms": 28418.166292016394, "ttft_ms": 13.62950005568564, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a47d2d9e37e8b4359e1c296f914ab9bec64c3ebe6d0c0381d50b8e0b27f8f7e2", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "cefca2be28e706dbc3f3ab03", "attempt_id": "e4f8fe0259da", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:35:50.679151Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 2070, "message_tokens": 1981, "message_utf8_bytes": 5176, "cache_tensor_bytes": null, "serialized_payload_bytes": 5176, "side_information_bytes": 0, "fetched_bytes": 2593, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 28260.889666969888, "ttft_ms": 13.519415981136262, "decode_ms": 96.59266599919647, "receiver_prefill_ms": 258.26116697862744}, "total_wall_ms": 28633.5017089732, "ttft_ms": 13.519415981136262, "input_usage": {"lookups": 9}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5236625", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "391110c477ecdfeafcdd5d5a16a124350470939f2762642d53e712d06c002690", "output_hash": "f0324cfa0e35a8d9bd4a75ef3f8cf62fcab822341c95b1d00f17b53fdf75ccf5"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "e97c22bdfca6cae9b944842e", "attempt_id": "32b4665da4f0", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:35:51.069944Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 1937, "message_tokens": 1845, "message_utf8_bytes": 4896, "cache_tensor_bytes": null, "serialized_payload_bytes": 4896, "side_information_bytes": 0, "fetched_bytes": 2319, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 28260.889666969888, "ttft_ms": 13.52625002618879, "decode_ms": 136.5687500219792, "receiver_prefill_ms": 205.8276670286432}, "total_wall_ms": 28620.994333992712, "ttft_ms": 13.52625002618879, "input_usage": {"lookups": 9}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "CN-53T-4564", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a1dba3598ea5ad5c2aa8e87229fc5abad0104e2bd5f6531dea9f98a2d6c983cf", "output_hash": "63b72c2f0fa9c87c99863654eaed185c2db37aca56aeb2c04043723bedec7f95"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "81f13aa3d23ee1559b878670", "attempt_id": "f53e6558a90a", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:35:51.448336Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 2055, "message_tokens": 1956, "message_utf8_bytes": 5155, "cache_tensor_bytes": null, "serialized_payload_bytes": 5155, "side_information_bytes": 0, "fetched_bytes": 2558, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 28260.889666969888, "ttft_ms": 13.445000047795475, "decode_ms": 13.500249944627285, "receiver_prefill_ms": 236.4125830354169}, "total_wall_ms": 28528.405708028004, "ttft_ms": 13.445000047795475, "input_usage": {"lookups": 9}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "d9f355034a4919b2ee63a2a6326aa36389eb4679525102e7afa500e0a0b642c9", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b6e36a040f319e262697019c", "attempt_id": "9a63ff3e7f6a", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:35:51.734949Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 2014, "message_tokens": 1924, "message_utf8_bytes": 5081, "cache_tensor_bytes": null, "serialized_payload_bytes": 5081, "side_information_bytes": 0, "fetched_bytes": 2497, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 28260.889666969888, "ttft_ms": 14.15095804259181, "decode_ms": 99.11412501242012, "receiver_prefill_ms": 318.8807499827817}, "total_wall_ms": 28695.990333915688, "ttft_ms": 14.15095804259181, "input_usage": {"lookups": 9}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5187820", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "3fe6371c6f66cd410891142d0caa565eb5f29914d319b00cc534e73498484005", "output_hash": "506e6115b34b564401e129c2ef42e2e7e0ec82b154b45d92c85846383969e6ed"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "30f41685451fb6668715f53a", "attempt_id": "690247df5fe2", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:35:52.188239Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 2001, "message_tokens": 1901, "message_utf8_bytes": 5043, "cache_tensor_bytes": null, "serialized_payload_bytes": 5043, "side_information_bytes": 0, "fetched_bytes": 2467, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 28260.889666969888, "ttft_ms": 13.637749943882227, "decode_ms": 67.8624160354957, "receiver_prefill_ms": 210.2904999628663}, "total_wall_ms": 28558.95545904059, "ttft_ms": 13.637749943882227, "input_usage": {"lookups": 9}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Urker-16", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "2d5b2bac54794fd20dbd44ef3a3e7c85045b6118ddda8d5e5b1884f61d8a8cb1", "output_hash": "13bcd50d62725e833c519fdb3eb450946e33fce2aca726f9c0f57cbe4cbcb2d9"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a067a602494a4009277e67f6", "attempt_id": "e76bb1c04b5f", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:35:52.505947Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 2049, "message_tokens": 1956, "message_utf8_bytes": 5155, "cache_tensor_bytes": null, "serialized_payload_bytes": 5155, "side_information_bytes": 0, "fetched_bytes": 2558, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 28260.889666969888, "ttft_ms": 14.416625024750829, "decode_ms": 110.2638749871403, "receiver_prefill_ms": 211.05158398859203}, "total_wall_ms": 28599.7276250273, "ttft_ms": 14.416625024750829, "input_usage": {"lookups": 9}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "14936855", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a9db72e148cbf9b5a96088fce4db88970b61c13aa21dc1be7fa74f9cf7adc24d", "output_hash": "c8d3235581060a2a3a6817385b337f6e5171e918972dd95b6188b095a1b4c66f"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "90ace0e0eb83ae351e2a0f37", "attempt_id": "7e9a6a4b4668", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:35:52.863082Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 2015, "message_tokens": 1922, "message_utf8_bytes": 5090, "cache_tensor_bytes": null, "serialized_payload_bytes": 5090, "side_information_bytes": 0, "fetched_bytes": 2483, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 28260.889666969888, "ttft_ms": 13.526208000257611, "decode_ms": 13.901083031669259, "receiver_prefill_ms": 218.87316601350904}, "total_wall_ms": 28510.188166983426, "ttft_ms": 13.526208000257611, "input_usage": {"lookups": 9}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "2542059506fb6e861bc185d1bc6e75874176fb6b21aea27149a6851d88a51e05", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "32723b898d953cad3c3bd983", "attempt_id": "0443c0183b65", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:35:53.130226Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 2051, "message_tokens": 1956, "message_utf8_bytes": 5155, "cache_tensor_bytes": null, "serialized_payload_bytes": 5155, "side_information_bytes": 0, "fetched_bytes": 2558, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 28260.889666969888, "ttft_ms": 13.819165993481874, "decode_ms": 13.771416968666017, "receiver_prefill_ms": 245.8000419428572}, "total_wall_ms": 28537.10575099103, "ttft_ms": 13.819165993481874, "input_usage": {"lookups": 9}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "d9f355034a4919b2ee63a2a6326aa36389eb4679525102e7afa500e0a0b642c9", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "80917d22fd23ef2d58ba9046", "attempt_id": "03211af3bc59", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:35:54.532270Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 238399488}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 43, "message_tokens": null, "message_utf8_bytes": 31201, "cache_tensor_bytes": 238399488, "serialized_payload_bytes": 238399488, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 1107.888541999273, "ttft_ms": 51.34804197587073, "decode_ms": 144.98691691551358, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1330.8715421007946, "ttft_ms": 51.34804197587073, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5,236,625", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "5f724075674359492cf09f6cc38fa29d1d8d57b1cbe36c12f18a44145249093b"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "ff6c4320ba71242276b2c170", "attempt_id": "41e00c0c523a", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:35:54.774512Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 238399488}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 46, "message_tokens": null, "message_utf8_bytes": 31201, "cache_tensor_bytes": 238399488, "serialized_payload_bytes": 238399488, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 1107.888541999273, "ttft_ms": 115.91662501450628, "decode_ms": 152.79374993406236, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1380.9211670886725, "ttft_ms": 115.91662501450628, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "CN-53T-4564", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "63b72c2f0fa9c87c99863654eaed185c2db37aca56aeb2c04043723bedec7f95"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "bac161ad34ac8ffe100db2cf", "attempt_id": "e4b76d646d61", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:35:55.066062Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 238399488}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 53, "message_tokens": null, "message_utf8_bytes": 31201, "cache_tensor_bytes": 238399488, "serialized_payload_bytes": 238399488, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 1107.888541999273, "ttft_ms": 137.10904109757394, "decode_ms": 106.36366705875844, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1355.7005421025679, "ttft_ms": 137.10904109757394, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7026040", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "745005bc5caf64de51f389cd215c24264b3c2b8761c288554f09868d28ea3ca3"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "f5952c08d2b5b6fd0a1c7e94", "attempt_id": "7c5970b0fc2a", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:35:55.332330Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 238399488}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 44, "message_tokens": null, "message_utf8_bytes": 31201, "cache_tensor_bytes": 238399488, "serialized_payload_bytes": 238399488, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 1107.888541999273, "ttft_ms": 108.21304202545434, "decode_ms": 136.78520801477134, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1357.512708986178, "ttft_ms": 108.21304202545434, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5,187,820", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "856804d5af6bf81ecc932fd68c47bd059046f4573ca64e21cb0c4c5d96dc59ca"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "5dd45c293a712bb722eccafa", "attempt_id": "77d38c1139a1", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:35:55.599373Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 238399488}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 54, "message_tokens": null, "message_utf8_bytes": 31201, "cache_tensor_bytes": 238399488, "serialized_payload_bytes": 238399488, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 1107.888541999273, "ttft_ms": 83.12412502709776, "decode_ms": 75.63141593709588, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1271.2796250125393, "ttft_ms": 83.12412502709776, "input_usage": {}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Urker-16", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "13bcd50d62725e833c519fdb3eb450946e33fce2aca726f9c0f57cbe4cbcb2d9"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "950d4a43d5b63381db725688", "attempt_id": "022c544fe97a", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:35:55.780625Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 238399488}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 31201, "cache_tensor_bytes": 238399488, "serialized_payload_bytes": 238399488, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 1107.888541999273, "ttft_ms": 48.36150002665818, "decode_ms": 121.27233296632767, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1282.0797079475597, "ttft_ms": 48.36150002665818, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "14213474", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "f66bfab08b868e35eea7b6d0e4f26c213c8255b6b13bef10cf210555ca814b1c"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "2f54baeafc8f993ec98d3688", "attempt_id": "bd33236c9e06", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:35:55.972759Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 238399488}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 31201, "cache_tensor_bytes": 238399488, "serialized_payload_bytes": 238399488, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 1107.888541999273, "ttft_ms": 48.36987506132573, "decode_ms": 15.249124960973859, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1175.6681670667604, "ttft_ms": 48.36987506132573, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "d49dad26792e0abab04f1120", "attempt_id": "80724acd095d", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:35:56.058112Z", "task_family": "synthetic", "document_id": "syn_03025_8000", "document_hash": "b2f6e5bf2101aa9e99d6aac4075349a1894a41e5e6b272f95bd8ed9f56300ad9", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03025_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 238399488}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5836, "query_tokens": 49, "message_tokens": null, "message_utf8_bytes": 31201, "cache_tensor_bytes": 238399488, "serialized_payload_bytes": 238399488, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 1107.888541999273, "ttft_ms": 47.45991597883403, "decode_ms": 15.515959006734192, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1175.4940008977428, "ttft_ms": 47.45991597883403, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Lunfel-18", "bogus_amount": "4,824,670", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03025_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "bc36bab5d59e6a2e990c66e5", "attempt_id": "3aa00ef117f4", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:35:57.096174Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 2407, "message_tokens": 2318, "message_utf8_bytes": 8288, "cache_tensor_bytes": null, "serialized_payload_bytes": 8288, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.033791991882026196, "ttft_ms": 13.843958033248782, "decode_ms": 96.99337498750538, "receiver_prefill_ms": 274.68375000171363}, "total_wall_ms": 391.58325106836855, "ttft_ms": 13.843958033248782, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "535,300", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "aa26eae05d558fd6e4c11a7079dbe3efe6574efb099211284d7f465a5c3027aa", "output_hash": "310282bc8b1d285db5e2b900f31e17e95e68f26b55e4eb87ed8d54f74105024b"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "793c7cf00dca25462bfe13c2", "attempt_id": "6a9b99598b23", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:35:57.501325Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 2404, "message_tokens": 2318, "message_utf8_bytes": 8288, "cache_tensor_bytes": null, "serialized_payload_bytes": 8288, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.033791991882026196, "ttft_ms": 13.730334001593292, "decode_ms": 135.3455410571769, "receiver_prefill_ms": 262.61141698341817}, "total_wall_ms": 415.3032510075718, "ttft_ms": 13.730334001593292, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "DB-72P-3503", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "aa26eae05d558fd6e4c11a7079dbe3efe6574efb099211284d7f465a5c3027aa", "output_hash": "051acaf42a5b0e2760fcb6fb7252c81cae4688e78f4f2cfa24c17e2daf75b9e3"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "37fda717b132068a0ac26b20", "attempt_id": "27b90baba243", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:35:57.930576Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 2411, "message_tokens": 2318, "message_utf8_bytes": 8288, "cache_tensor_bytes": null, "serialized_payload_bytes": 8288, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.033791991882026196, "ttft_ms": 13.621708028949797, "decode_ms": 95.34083295147866, "receiver_prefill_ms": 263.4287910768762}, "total_wall_ms": 376.0849590180442, "ttft_ms": 13.621708028949797, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8010205", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "aa26eae05d558fd6e4c11a7079dbe3efe6574efb099211284d7f465a5c3027aa", "output_hash": "6d1ceb36132f38bd77e8f46c23ea21dd9bcff78bc80bd8a9eaf2d7f2f3bacf46"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "8bf3b2638ea207f74275fa64", "attempt_id": "cc73ecd9d01c", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:35:58.320496Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 2404, "message_tokens": 2318, "message_utf8_bytes": 8288, "cache_tensor_bytes": null, "serialized_payload_bytes": 8288, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.033791991882026196, "ttft_ms": 13.734625070355833, "decode_ms": 122.26495903450996, "receiver_prefill_ms": 263.12816701829433}, "total_wall_ms": 402.7411669958383, "ttft_ms": 13.734625070355833, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5,632,215", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "aa26eae05d558fd6e4c11a7079dbe3efe6574efb099211284d7f465a5c3027aa", "output_hash": "eda75ec7ef30171e592506abc3e5b148b641b914854fc239672772c94e5c3232"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "f9e09c202826803ea840fc91", "attempt_id": "78bc89a3e947", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:35:58.737209Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 2414, "message_tokens": 2318, "message_utf8_bytes": 8288, "cache_tensor_bytes": null, "serialized_payload_bytes": 8288, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.033791991882026196, "ttft_ms": 13.50137498229742, "decode_ms": 95.38145805709064, "receiver_prefill_ms": 263.3870409335941}, "total_wall_ms": 375.90137496590614, "ttft_ms": 13.50137498229742, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Lunlith-13", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "aa26eae05d558fd6e4c11a7079dbe3efe6574efb099211284d7f465a5c3027aa", "output_hash": "c45f000afe6d865c8167954eebdff761141225817d4f5b26c962cd50cd75d391"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "2e04d7fc68c9a1411acd4d07", "attempt_id": "9a8a0cc4b84b", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:35:59.126973Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 2407, "message_tokens": 2318, "message_utf8_bytes": 8288, "cache_tensor_bytes": null, "serialized_payload_bytes": 8288, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.033791991882026196, "ttft_ms": 13.621708028949797, "decode_ms": 94.62620900012553, "receiver_prefill_ms": 263.62941693514585}, "total_wall_ms": 375.56279194541276, "ttft_ms": 13.621708028949797, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "2516345", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "aa26eae05d558fd6e4c11a7079dbe3efe6574efb099211284d7f465a5c3027aa", "output_hash": "c9746d6b321a85e7860df5d1300d2968d4ee3af4bbfa9b3c78c9484900e77381"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "4fc8e6ff48051f154aaa1751", "attempt_id": "f8843076f5ff", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:35:59.516399Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 2407, "message_tokens": 2318, "message_utf8_bytes": 8288, "cache_tensor_bytes": null, "serialized_payload_bytes": 8288, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.033791991882026196, "ttft_ms": 13.392583001405, "decode_ms": 13.639791985042393, "receiver_prefill_ms": 262.9818330751732}, "total_wall_ms": 293.57912589330226, "ttft_ms": 13.392583001405, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "aa26eae05d558fd6e4c11a7079dbe3efe6574efb099211284d7f465a5c3027aa", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "7ba617bbaad3f5a63a941dd5", "attempt_id": "a80d04f906cb", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:35:59.824590Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 2409, "message_tokens": 2318, "message_utf8_bytes": 8288, "cache_tensor_bytes": null, "serialized_payload_bytes": 8288, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.033791991882026196, "ttft_ms": 13.725041993893683, "decode_ms": 13.780791079625487, "receiver_prefill_ms": 264.4435829715803}, "total_wall_ms": 295.5767089733854, "ttft_ms": 13.725041993893683, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "aa26eae05d558fd6e4c11a7079dbe3efe6574efb099211284d7f465a5c3027aa", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "28e069a34d07ff24535d2afc", "attempt_id": "18f0ceb9b1c7", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:36:14.218349Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 1127, "message_tokens": 1036, "message_utf8_bytes": 2369, "cache_tensor_bytes": null, "serialized_payload_bytes": 2369, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14084.534749970771, "ttft_ms": 13.7305409880355, "decode_ms": 94.44662497844547, "receiver_prefill_ms": 121.7910829000175}, "total_wall_ms": 14316.346333012916, "ttft_ms": 13.7305409880355, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "535,300", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "c5638ab9853837d89a445e697804f09f6a3d0cf6c2a0fb955da5b8d8010c988f", "output_hash": "310282bc8b1d285db5e2b900f31e17e95e68f26b55e4eb87ed8d54f74105024b"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "628e7ff7bbd8f13f13995e0f", "attempt_id": "107c3df6f4c0", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:36:14.464284Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 1124, "message_tokens": 1036, "message_utf8_bytes": 2369, "cache_tensor_bytes": null, "serialized_payload_bytes": 2369, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14084.534749970771, "ttft_ms": 13.400291092693806, "decode_ms": 135.31916600186378, "receiver_prefill_ms": 120.95191597472876}, "total_wall_ms": 14356.111874920316, "ttft_ms": 13.400291092693806, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "DB-72P-3503", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "c5638ab9853837d89a445e697804f09f6a3d0cf6c2a0fb955da5b8d8010c988f", "output_hash": "051acaf42a5b0e2760fcb6fb7252c81cae4688e78f4f2cfa24c17e2daf75b9e3"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b209ab8db43e690c3b6c5f22", "attempt_id": "4cffe31ff6e2", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:36:14.749902Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 1131, "message_tokens": 1036, "message_utf8_bytes": 2369, "cache_tensor_bytes": null, "serialized_payload_bytes": 2369, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14084.534749970771, "ttft_ms": 13.88458302244544, "decode_ms": 94.98499997425824, "receiver_prefill_ms": 121.31541699636728}, "total_wall_ms": 14316.582540981472, "ttft_ms": 13.88458302244544, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8010205", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "c5638ab9853837d89a445e697804f09f6a3d0cf6c2a0fb955da5b8d8010c988f", "output_hash": "6d1ceb36132f38bd77e8f46c23ea21dd9bcff78bc80bd8a9eaf2d7f2f3bacf46"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a9455032c10aa4d84f764bc2", "attempt_id": "5fbf28bda231", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:36:14.996035Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 1124, "message_tokens": 1036, "message_utf8_bytes": 2369, "cache_tensor_bytes": null, "serialized_payload_bytes": 2369, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14084.534749970771, "ttft_ms": 13.530290918424726, "decode_ms": 122.91462498251349, "receiver_prefill_ms": 122.15995800215751}, "total_wall_ms": 14345.030582975596, "ttft_ms": 13.530290918424726, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "6,056,150", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "c5638ab9853837d89a445e697804f09f6a3d0cf6c2a0fb955da5b8d8010c988f", "output_hash": "8262889a8478c7e0ee965f159d9849a3849b412e276f5b979476e2da5b320249"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "806e2b0b05c9792f96b0a70f", "attempt_id": "72f4720c3742", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:36:15.272032Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 1134, "message_tokens": 1036, "message_utf8_bytes": 2369, "cache_tensor_bytes": null, "serialized_payload_bytes": 2369, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14084.534749970771, "ttft_ms": 13.966791098937392, "decode_ms": 68.20216693449765, "receiver_prefill_ms": 123.67983302101493}, "total_wall_ms": 14295.083042001352, "ttft_ms": 13.966791098937392, "input_usage": {}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Novost-15", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "c5638ab9853837d89a445e697804f09f6a3d0cf6c2a0fb955da5b8d8010c988f", "output_hash": "851a475fb1ba6ddb5b9988bd32045e7b0c809becb5edebc929cce56a70ed5a5d"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "6dbe8579ecc50c7cf5fddeaf", "attempt_id": "c93a1c749d87", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:36:15.496849Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 1127, "message_tokens": 1036, "message_utf8_bytes": 2369, "cache_tensor_bytes": null, "serialized_payload_bytes": 2369, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14084.534749970771, "ttft_ms": 13.531415956094861, "decode_ms": 110.03220803104341, "receiver_prefill_ms": 121.71845801640302}, "total_wall_ms": 14331.66349993553, "ttft_ms": 13.531415956094861, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "15173162", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "c5638ab9853837d89a445e697804f09f6a3d0cf6c2a0fb955da5b8d8010c988f", "output_hash": "cc03b0e0d18e28d4b86c050a4e72102728582e4c3eddce83767093d208d86863"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "09d0fc90ae2d9f4b69117c11", "attempt_id": "cc3a0e876434", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:36:15.758381Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 1127, "message_tokens": 1036, "message_utf8_bytes": 2369, "cache_tensor_bytes": null, "serialized_payload_bytes": 2369, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14084.534749970771, "ttft_ms": 13.2902089972049, "decode_ms": 13.595665921457112, "receiver_prefill_ms": 121.89254094846547}, "total_wall_ms": 14235.049540991895, "ttft_ms": 13.2902089972049, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "c5638ab9853837d89a445e697804f09f6a3d0cf6c2a0fb955da5b8d8010c988f", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "549fd0e7043a4b6b89e097c9", "attempt_id": "2a6f64a665d5", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:36:15.923222Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 1129, "message_tokens": 1036, "message_utf8_bytes": 2369, "cache_tensor_bytes": null, "serialized_payload_bytes": 2369, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14084.534749970771, "ttft_ms": 13.373833964578807, "decode_ms": 13.39112501591444, "receiver_prefill_ms": 120.9158330457285}, "total_wall_ms": 14234.04383403249, "ttft_ms": 13.373833964578807, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "c5638ab9853837d89a445e697804f09f6a3d0cf6c2a0fb955da5b8d8010c988f", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "89b632f0e6de81c95ea0c962", "attempt_id": "3fb87ff1f110", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:36:30.196956Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 14}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 641, "message_tokens": 547, "message_utf8_bytes": 1433, "cache_tensor_bytes": null, "serialized_payload_bytes": 1433, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14110.15425005462, "ttft_ms": 13.454999891109765, "decode_ms": 13.339249999262393, "receiver_prefill_ms": 72.18979101162404}, "total_wall_ms": 14215.450625051744, "ttft_ms": 13.454999891109765, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "26e8142bbe991631989d1ca5c21214dedae6fd0f796dae52a765cde97c9deae5", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b86e4e04a0f5eec08cba29d7", "attempt_id": "ae338b2ec1ae", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:36:30.316294Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 14}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 638, "message_tokens": 547, "message_utf8_bytes": 1433, "cache_tensor_bytes": null, "serialized_payload_bytes": 1433, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14110.15425005462, "ttft_ms": 13.252333970740438, "decode_ms": 133.0637080827728, "receiver_prefill_ms": 73.46387498546392}, "total_wall_ms": 14331.303874962032, "ttft_ms": 13.252333970740438, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "WT-26M-9227", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "26e8142bbe991631989d1ca5c21214dedae6fd0f796dae52a765cde97c9deae5", "output_hash": "ef2302e0ab6f5a6951b11463fea177ba766b9d438afcaf6e6101ebfa3ba38aba"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "692e7b9f9202efa07afc77f4", "attempt_id": "89fa9d45600e", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:36:30.551484Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 14}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 645, "message_tokens": 547, "message_utf8_bytes": 1433, "cache_tensor_bytes": null, "serialized_payload_bytes": 1433, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14110.15425005462, "ttft_ms": 13.430458959192038, "decode_ms": 93.03074993658811, "receiver_prefill_ms": 75.07470808923244}, "total_wall_ms": 14292.996709118597, "ttft_ms": 13.430458959192038, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "6914937", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "26e8142bbe991631989d1ca5c21214dedae6fd0f796dae52a765cde97c9deae5", "output_hash": "534d40ded32f784c68307dd2386faed2a27bf39dcc27321ee6985e42ccd285e5"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b369df86dc6537166b4887e4", "attempt_id": "aecf818a2195", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:36:30.748671Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 14}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 638, "message_tokens": 547, "message_utf8_bytes": 1433, "cache_tensor_bytes": null, "serialized_payload_bytes": 1433, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14110.15425005462, "ttft_ms": 13.226458104327321, "decode_ms": 94.4082080386579, "receiver_prefill_ms": 73.31841601990163}, "total_wall_ms": 14292.383708059788, "ttft_ms": 13.226458104327321, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "6914937", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "26e8142bbe991631989d1ca5c21214dedae6fd0f796dae52a765cde97c9deae5", "output_hash": "534d40ded32f784c68307dd2386faed2a27bf39dcc27321ee6985e42ccd285e5"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "39e4d22bb8406446118140c0", "attempt_id": "2a04e53e9744", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:36:30.945118Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 14}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 648, "message_tokens": 547, "message_utf8_bytes": 1433, "cache_tensor_bytes": null, "serialized_payload_bytes": 1433, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14110.15425005462, "ttft_ms": 13.290166039951146, "decode_ms": 13.39766604360193, "receiver_prefill_ms": 75.19283390138298}, "total_wall_ms": 14213.356374995783, "ttft_ms": 13.290166039951146, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "26e8142bbe991631989d1ca5c21214dedae6fd0f796dae52a765cde97c9deae5", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c8ec9682cb430a969829f401", "attempt_id": "2f2ee585a33d", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:36:31.062632Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 14}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 641, "message_tokens": 547, "message_utf8_bytes": 1433, "cache_tensor_bytes": null, "serialized_payload_bytes": 1433, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14110.15425005462, "ttft_ms": 13.828041031956673, "decode_ms": 132.66812497749925, "receiver_prefill_ms": 71.70620793476701}, "total_wall_ms": 14329.704083036631, "ttft_ms": 13.828041031956673, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "10,244,447", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "26e8142bbe991631989d1ca5c21214dedae6fd0f796dae52a765cde97c9deae5", "output_hash": "041c4a94218fbbc63e1da6dced8a0ea330f7750c6d93cf330eee091c4936494a"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "78b57e197b8e77576af932f8", "attempt_id": "16b99d88724c", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:36:31.295912Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 14}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 641, "message_tokens": 547, "message_utf8_bytes": 1433, "cache_tensor_bytes": null, "serialized_payload_bytes": 1433, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14110.15425005462, "ttft_ms": 13.43870791606605, "decode_ms": 13.248666073195636, "receiver_prefill_ms": 72.09612498991191}, "total_wall_ms": 14210.346875130199, "ttft_ms": 13.43870791606605, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "26e8142bbe991631989d1ca5c21214dedae6fd0f796dae52a765cde97c9deae5", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "63825ac5beb334c1cbb53f15", "attempt_id": "ed1a3639f5a6", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:36:31.410060Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 14}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 643, "message_tokens": 547, "message_utf8_bytes": 1433, "cache_tensor_bytes": null, "serialized_payload_bytes": 1433, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14110.15425005462, "ttft_ms": 13.408041908405721, "decode_ms": 13.291583978570998, "receiver_prefill_ms": 74.8914999421686}, "total_wall_ms": 14213.045833050273, "ttft_ms": 13.408041908405721, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "26e8142bbe991631989d1ca5c21214dedae6fd0f796dae52a765cde97c9deae5", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "4f407ebbe18bda1b5b1ce0ae", "attempt_id": "d835e43a35cf", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:36:45.758493Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 1415, "message_tokens": 1322, "message_utf8_bytes": 3365, "cache_tensor_bytes": null, "serialized_payload_bytes": 3365, "side_information_bytes": 0, "fetched_bytes": 1792, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14231.364666949958, "ttft_ms": 13.874458032660186, "decode_ms": 13.801709050312638, "receiver_prefill_ms": 151.47399995476007}, "total_wall_ms": 14412.727791932411, "ttft_ms": 13.874458032660186, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "f17b17688d45b7f39f8f52663a70e30a22a28efb06855699acc9e867548eb80e", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "48a92f0630f7b8814eb3e10a", "attempt_id": "39654f8d9884", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:36:45.954619Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 1391, "message_tokens": 1301, "message_utf8_bytes": 3302, "cache_tensor_bytes": null, "serialized_payload_bytes": 3302, "side_information_bytes": 0, "fetched_bytes": 1727, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14231.364666949958, "ttft_ms": 13.841332984156907, "decode_ms": 55.17562502063811, "receiver_prefill_ms": 148.6471249954775}, "total_wall_ms": 14451.240957947448, "ttft_ms": 13.841332984156907, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 4}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "DB-72", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "c3c6b66ff1f2922d358e41bfd7567b290f3cf4c0308ebf80ae859f3c5269d953", "output_hash": "87ca83e01929be6b3f8de6994a2e6b49ab938a1a8dd15ee4968880f98d5f92b7"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "2ab0a340ed478b1dc1be9f6b", "attempt_id": "b2ccec9ed407", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:36:46.189466Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 1405, "message_tokens": 1308, "message_utf8_bytes": 3319, "cache_tensor_bytes": null, "serialized_payload_bytes": 3319, "side_information_bytes": 0, "fetched_bytes": 1745, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14231.364666949958, "ttft_ms": 13.553750002756715, "decode_ms": 97.26200008299202, "receiver_prefill_ms": 149.11612507421523}, "total_wall_ms": 14493.502416997217, "ttft_ms": 13.553750002756715, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8010205", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "90077163f961765236ec89339f3613d11e838181e6b88cb141a2f8f68cd2d8ff", "output_hash": "6d1ceb36132f38bd77e8f46c23ea21dd9bcff78bc80bd8a9eaf2d7f2f3bacf46"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "fbd63fdeeea75c08c5841b8a", "attempt_id": "8154bbeea164", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:36:46.466023Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 1398, "message_tokens": 1308, "message_utf8_bytes": 3319, "cache_tensor_bytes": null, "serialized_payload_bytes": 3319, "side_information_bytes": 0, "fetched_bytes": 1745, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14231.364666949958, "ttft_ms": 13.884167070500553, "decode_ms": 97.9517080122605, "receiver_prefill_ms": 148.31058296840638}, "total_wall_ms": 14493.840291863307, "ttft_ms": 13.884167070500553, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "6056150", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "90077163f961765236ec89339f3613d11e838181e6b88cb141a2f8f68cd2d8ff", "output_hash": "f016f8b5fc26708e6c1ff75ff8e98659082c41a168f273819ac48b88feea8d3f"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "69e8009cf7273bdb1570d974", "attempt_id": "de6473f63c4a", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:36:46.743403Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 1368, "message_tokens": 1268, "message_utf8_bytes": 3312, "cache_tensor_bytes": null, "serialized_payload_bytes": 3312, "side_information_bytes": 0, "fetched_bytes": 1735, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14231.364666949958, "ttft_ms": 13.741832925006747, "decode_ms": 68.19629098754376, "receiver_prefill_ms": 144.84883390832692}, "total_wall_ms": 14460.509749944322, "ttft_ms": 13.741832925006747, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Novost-15", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "f75065b9ea599e2e1d5c96b191ccd822e54a59878a7ac9e589f9f7eee0ee1c5f", "output_hash": "851a475fb1ba6ddb5b9988bd32045e7b0c809becb5edebc929cce56a70ed5a5d"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "de13ffeccac4b5acc23706b9", "attempt_id": "42a0a9b29014", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:36:46.987146Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 1415, "message_tokens": 1322, "message_utf8_bytes": 3365, "cache_tensor_bytes": null, "serialized_payload_bytes": 3365, "side_information_bytes": 0, "fetched_bytes": 1792, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14231.364666949958, "ttft_ms": 13.382208999246359, "decode_ms": 95.89595894794911, "receiver_prefill_ms": 150.43208305723965}, "total_wall_ms": 14493.294833926484, "ttft_ms": 13.382208999246359, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "6914937", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "4fb1d2eac13c20331f2366348f886965516aa167cec0d5326617d0dfa8a0cf2e", "output_hash": "534d40ded32f784c68307dd2386faed2a27bf39dcc27321ee6985e42ccd285e5"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "e9bb231186ecce77ea340f73", "attempt_id": "b0fefbdc209e", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:36:47.264346Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 1394, "message_tokens": 1301, "message_utf8_bytes": 3302, "cache_tensor_bytes": null, "serialized_payload_bytes": 3302, "side_information_bytes": 0, "fetched_bytes": 1727, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14231.364666949958, "ttft_ms": 13.66433291696012, "decode_ms": 13.789082993753254, "receiver_prefill_ms": 147.96349999960512}, "total_wall_ms": 14409.027416957542, "ttft_ms": 13.66433291696012, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "c3c6b66ff1f2922d358e41bfd7567b290f3cf4c0308ebf80ae859f3c5269d953", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a637e5af49e6df1942451e2b", "attempt_id": "f786dacfece8", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:36:47.456153Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 1410, "message_tokens": 1315, "message_utf8_bytes": 3318, "cache_tensor_bytes": null, "serialized_payload_bytes": 3318, "side_information_bytes": 0, "fetched_bytes": 1745, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14231.364666949958, "ttft_ms": 13.809291063807905, "decode_ms": 13.91787501052022, "receiver_prefill_ms": 150.78795899171382}, "total_wall_ms": 14412.027666927315, "ttft_ms": 13.809291063807905, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "7be52cf7a301039a35da20ab166eb1fac7e8a0c2625d4c50f876dcd882834df8", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b3edeaa43655f4bcf1188629", "attempt_id": "8663fa9c49b7", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:36:47.935575Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 88326144}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 8262, "cache_tensor_bytes": 88326144, "serialized_payload_bytes": 88326144, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 284.1506660915911, "ttft_ms": 31.503833015449345, "decode_ms": 99.41104205790907, "receiver_prefill_ms": 0.0}, "total_wall_ms": 419.49191608000547, "ttft_ms": 31.503833015449345, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "535,300", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "310282bc8b1d285db5e2b900f31e17e95e68f26b55e4eb87ed8d54f74105024b"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c4fa69d89078ea2cd709e4c1", "attempt_id": "113dc25c1021", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:36:48.085148Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 88326144}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 44, "message_tokens": null, "message_utf8_bytes": 8262, "cache_tensor_bytes": 88326144, "serialized_payload_bytes": 88326144, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 284.1506660915911, "ttft_ms": 29.02508294209838, "decode_ms": 140.98566700704396, "receiver_prefill_ms": 0.0}, "total_wall_ms": 458.03991600405425, "ttft_ms": 29.02508294209838, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "DD-69N-1957", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "c7a79bc9b8b76dec6be02055bce804a678265ed054fadbc587c4eb56d55e06dd"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b9aacd7585e3c1172652901d", "attempt_id": "699c4fc7b7a5", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:36:48.273104Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 88326144}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 51, "message_tokens": null, "message_utf8_bytes": 8262, "cache_tensor_bytes": 88326144, "serialized_payload_bytes": 88326144, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 284.1506660915911, "ttft_ms": 29.24274990800768, "decode_ms": 98.2317499583587, "receiver_prefill_ms": 0.0}, "total_wall_ms": 415.44412507209927, "ttft_ms": 29.24274990800768, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5632215", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "dff2a8edf1f157947cce1a41ebc57f74d76214f6c6a3ee4b8447fd1333253dba"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "1bb23e54972bada488d91a3d", "attempt_id": "156cb0f9ee5e", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:36:48.418306Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 88326144}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 44, "message_tokens": null, "message_utf8_bytes": 8262, "cache_tensor_bytes": 88326144, "serialized_payload_bytes": 88326144, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 284.1506660915911, "ttft_ms": 27.50070800539106, "decode_ms": 99.4282920146361, "receiver_prefill_ms": 0.0}, "total_wall_ms": 414.89820706192404, "ttft_ms": 27.50070800539106, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5632215", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "dff2a8edf1f157947cce1a41ebc57f74d76214f6c6a3ee4b8447fd1333253dba"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "088cb96ce3df93244ef947c6", "attempt_id": "9a0720ff1096", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:36:48.562719Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 88326144}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 54, "message_tokens": null, "message_utf8_bytes": 8262, "cache_tensor_bytes": 88326144, "serialized_payload_bytes": 88326144, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 284.1506660915911, "ttft_ms": 28.828584006987512, "decode_ms": 99.24116602633148, "receiver_prefill_ms": 0.0}, "total_wall_ms": 416.13904119003564, "ttft_ms": 28.828584006987512, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Lunlith-13", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "c45f000afe6d865c8167954eebdff761141225817d4f5b26c962cd50cd75d391"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "972269857842b5f0c002da6c", "attempt_id": "ad901951db5c", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:36:48.708735Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 88326144}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 8262, "cache_tensor_bytes": 88326144, "serialized_payload_bytes": 88326144, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 284.1506660915911, "ttft_ms": 29.08729191403836, "decode_ms": 112.94387502130121, "receiver_prefill_ms": 0.0}, "total_wall_ms": 430.3772911662236, "ttft_ms": 29.08729191403836, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "14435552", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "ba9c717ba0089d9eed09e98d01b9e69c0f54309241850f2266950a76789eaba3"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "6332ff2e9e2ffe0b20f64a1e", "attempt_id": "9514773e3c4b", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:36:48.870348Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 88326144}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 8262, "cache_tensor_bytes": 88326144, "serialized_payload_bytes": 88326144, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 284.1506660915911, "ttft_ms": 28.757916996255517, "decode_ms": 14.211958041414618, "receiver_prefill_ms": 0.0}, "total_wall_ms": 330.1147910533473, "ttft_ms": 28.757916996255517, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "646ddccc6fd5381444ae2368", "attempt_id": "8bead021c19f", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:36:48.931165Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 88326144}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 49, "message_tokens": null, "message_utf8_bytes": 8262, "cache_tensor_bytes": 88326144, "serialized_payload_bytes": 88326144, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 284.1506660915911, "ttft_ms": 29.64787499513477, "decode_ms": 14.123957953415811, "receiver_prefill_ms": 0.0}, "total_wall_ms": 331.91920816898346, "ttft_ms": 29.64787499513477, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "76ccb4b0bd3bcb71385cb0b0", "attempt_id": "8ebc0e61fa73", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:36:48.993543Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 2521, "message_tokens": 2432, "message_utf8_bytes": 8792, "cache_tensor_bytes": null, "serialized_payload_bytes": 8792, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.026000081561505795, "ttft_ms": 13.747415971010923, "decode_ms": 96.7805830296129, "receiver_prefill_ms": 272.9843750130385}, "total_wall_ms": 387.3797501437366, "ttft_ms": 13.747415971010923, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "535,300", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "b90b8ab54a04ff718678281d382d7e5b0e7dd79c8d0bdadc63dbaade42384509", "output_hash": "310282bc8b1d285db5e2b900f31e17e95e68f26b55e4eb87ed8d54f74105024b"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "2144841f00d0a06a798e7d71", "attempt_id": "ff821251f03c", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:36:49.394577Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 2518, "message_tokens": 2432, "message_utf8_bytes": 8792, "cache_tensor_bytes": null, "serialized_payload_bytes": 8792, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.026000081561505795, "ttft_ms": 13.569499948062003, "decode_ms": 136.36124995537102, "receiver_prefill_ms": 271.18412498384714}, "total_wall_ms": 424.92104205302894, "ttft_ms": 13.569499948062003, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "DB-72P-3503", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "b90b8ab54a04ff718678281d382d7e5b0e7dd79c8d0bdadc63dbaade42384509", "output_hash": "051acaf42a5b0e2760fcb6fb7252c81cae4688e78f4f2cfa24c17e2daf75b9e3"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "97571e9de1ca6421edcb2bc3", "attempt_id": "09ceff6c4728", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:36:49.833552Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 2525, "message_tokens": 2432, "message_utf8_bytes": 8792, "cache_tensor_bytes": null, "serialized_payload_bytes": 8792, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.026000081561505795, "ttft_ms": 13.496999978087842, "decode_ms": 95.73508403263986, "receiver_prefill_ms": 272.04050007276237}, "total_wall_ms": 385.08800009731203, "ttft_ms": 13.496999978087842, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8010205", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "b90b8ab54a04ff718678281d382d7e5b0e7dd79c8d0bdadc63dbaade42384509", "output_hash": "6d1ceb36132f38bd77e8f46c23ea21dd9bcff78bc80bd8a9eaf2d7f2f3bacf46"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "722663b2dcf54addeb3bb35d", "attempt_id": "2f8d108e9f03", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:36:50.233024Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 2518, "message_tokens": 2432, "message_utf8_bytes": 8792, "cache_tensor_bytes": null, "serialized_payload_bytes": 8792, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.026000081561505795, "ttft_ms": 13.704334036447108, "decode_ms": 123.62491700332612, "receiver_prefill_ms": 271.5251670451835}, "total_wall_ms": 412.6498340629041, "ttft_ms": 13.704334036447108, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5,632,215", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "b90b8ab54a04ff718678281d382d7e5b0e7dd79c8d0bdadc63dbaade42384509", "output_hash": "eda75ec7ef30171e592506abc3e5b148b641b914854fc239672772c94e5c3232"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "912b9b05ed37727d4eb13368", "attempt_id": "1268fcd2b872", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:36:50.659505Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 2528, "message_tokens": 2432, "message_utf8_bytes": 8792, "cache_tensor_bytes": null, "serialized_payload_bytes": 8792, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.026000081561505795, "ttft_ms": 13.81883304566145, "decode_ms": 69.4627920165658, "receiver_prefill_ms": 293.484624940902}, "total_wall_ms": 380.7400841033086, "ttft_ms": 13.81883304566145, "input_usage": {}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Novost-15", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "b90b8ab54a04ff718678281d382d7e5b0e7dd79c8d0bdadc63dbaade42384509", "output_hash": "851a475fb1ba6ddb5b9988bd32045e7b0c809becb5edebc929cce56a70ed5a5d"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b2201bdd2ab6ec54e4df2796", "attempt_id": "10ca17af350f", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:36:51.054653Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 2521, "message_tokens": 2432, "message_utf8_bytes": 8792, "cache_tensor_bytes": null, "serialized_payload_bytes": 8792, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.026000081561505795, "ttft_ms": 13.545709080062807, "decode_ms": 95.07220797240734, "receiver_prefill_ms": 270.9655000362545}, "total_wall_ms": 383.55291611514986, "ttft_ms": 13.545709080062807, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "2347846", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "b90b8ab54a04ff718678281d382d7e5b0e7dd79c8d0bdadc63dbaade42384509", "output_hash": "42c74d29c7a4ddecc47d82b3608b9baf822d4f1b8fd2ab248d6fcb503656e384"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "e603126fe102e5239e6d7df4", "attempt_id": "8144b926684c", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:36:51.452883Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 2521, "message_tokens": 2432, "message_utf8_bytes": 8792, "cache_tensor_bytes": null, "serialized_payload_bytes": 8792, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.026000081561505795, "ttft_ms": 13.71395809110254, "decode_ms": 13.348375097848475, "receiver_prefill_ms": 271.22762496583164}, "total_wall_ms": 302.076542051509, "ttft_ms": 13.71395809110254, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "b90b8ab54a04ff718678281d382d7e5b0e7dd79c8d0bdadc63dbaade42384509", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "8de3046a14b6d3acb4bbcbca", "attempt_id": "00c662944437", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:36:51.768780Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 2523, "message_tokens": 2432, "message_utf8_bytes": 8792, "cache_tensor_bytes": null, "serialized_payload_bytes": 8792, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.026000081561505795, "ttft_ms": 13.700208044610918, "decode_ms": 13.669624924659729, "receiver_prefill_ms": 277.95716701075435}, "total_wall_ms": 309.1204169904813, "ttft_ms": 13.700208044610918, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "b90b8ab54a04ff718678281d382d7e5b0e7dd79c8d0bdadc63dbaade42384509", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "3e0b6aaba1663ee7e8cb9cbd", "attempt_id": "98a31938f68a", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:37:12.618786Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1500}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 1614, "message_tokens": 1524, "message_utf8_bytes": 4770, "cache_tensor_bytes": null, "serialized_payload_bytes": 4770, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 20526.26595797483, "ttft_ms": 13.802000088617206, "decode_ms": 95.43770900927484, "receiver_prefill_ms": 171.59250006079674}, "total_wall_ms": 20812.575957970694, "ttft_ms": 13.802000088617206, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "535,300", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ecf5a44f55b5a0c2cfb61d3be37f3d239f5f933dcd1227578cf96f16f839a7be", "output_hash": "310282bc8b1d285db5e2b900f31e17e95e68f26b55e4eb87ed8d54f74105024b"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c9150d055855ce701643649e", "attempt_id": "3caab6d8a381", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:37:12.920182Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1500}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 1611, "message_tokens": 1524, "message_utf8_bytes": 4770, "cache_tensor_bytes": null, "serialized_payload_bytes": 4770, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 20526.26595797483, "ttft_ms": 13.629000051878393, "decode_ms": 134.59487492218614, "receiver_prefill_ms": 170.70274997968227}, "total_wall_ms": 20847.8197499644, "ttft_ms": 13.629000051878393, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "DB-72P-3503", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ecf5a44f55b5a0c2cfb61d3be37f3d239f5f933dcd1227578cf96f16f839a7be", "output_hash": "051acaf42a5b0e2760fcb6fb7252c81cae4688e78f4f2cfa24c17e2daf75b9e3"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a52d968fca241b9a6f6141c0", "attempt_id": "66579940a204", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:37:13.256177Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1500}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 1618, "message_tokens": 1524, "message_utf8_bytes": 4770, "cache_tensor_bytes": null, "serialized_payload_bytes": 4770, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 20526.26595797483, "ttft_ms": 13.731667073443532, "decode_ms": 96.73404204659164, "receiver_prefill_ms": 171.3238749653101}, "total_wall_ms": 20810.726124909706, "ttft_ms": 13.731667073443532, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5632215", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ecf5a44f55b5a0c2cfb61d3be37f3d239f5f933dcd1227578cf96f16f839a7be", "output_hash": "dff2a8edf1f157947cce1a41ebc57f74d76214f6c6a3ee4b8447fd1333253dba"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "7f1245ef3d4fe7ed2eeace6f", "attempt_id": "988a2dd96ed6", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:37:13.555303Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1500}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 1611, "message_tokens": 1524, "message_utf8_bytes": 4770, "cache_tensor_bytes": null, "serialized_payload_bytes": 4770, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 20526.26595797483, "ttft_ms": 13.817999977618456, "decode_ms": 122.9941250057891, "receiver_prefill_ms": 170.25370802730322}, "total_wall_ms": 20835.99662396591, "ttft_ms": 13.817999977618456, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "6,914,937", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ecf5a44f55b5a0c2cfb61d3be37f3d239f5f933dcd1227578cf96f16f839a7be", "output_hash": "c56851abfdbf8851226a1e71c4911c616a691c490a883eb754e61bac551d004c"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "319d40d4e52c17a967560258", "attempt_id": "406233935142", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:37:13.880573Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1500}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 1621, "message_tokens": 1524, "message_utf8_bytes": 4770, "cache_tensor_bytes": null, "serialized_payload_bytes": 4770, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 20526.26595797483, "ttft_ms": 13.538833940401673, "decode_ms": 67.90708296466619, "receiver_prefill_ms": 170.2840420184657}, "total_wall_ms": 20780.672374996357, "ttft_ms": 13.538833940401673, "input_usage": {}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Novost-15", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ecf5a44f55b5a0c2cfb61d3be37f3d239f5f933dcd1227578cf96f16f839a7be", "output_hash": "851a475fb1ba6ddb5b9988bd32045e7b0c809becb5edebc929cce56a70ed5a5d"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "04decc11ef704ba043fa516d", "attempt_id": "357c45023461", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:37:14.150603Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1500}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 1614, "message_tokens": 1524, "message_utf8_bytes": 4770, "cache_tensor_bytes": null, "serialized_payload_bytes": 4770, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 20526.26595797483, "ttft_ms": 13.540958985686302, "decode_ms": 110.25925003923476, "receiver_prefill_ms": 170.5873750615865}, "total_wall_ms": 20823.310791049153, "ttft_ms": 13.540958985686302, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "15233452", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ecf5a44f55b5a0c2cfb61d3be37f3d239f5f933dcd1227578cf96f16f839a7be", "output_hash": "946ed57d0fc8dad61733bd4e020c2c84a76dd92fce748c4069c1d2fd0c982b96"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "2d36e9d22df210bc58b6c032", "attempt_id": "a75595a9dd2e", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:37:14.462245Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1500}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 1614, "message_tokens": 1524, "message_utf8_bytes": 4770, "cache_tensor_bytes": null, "serialized_payload_bytes": 4770, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 20526.26595797483, "ttft_ms": 13.848332921043038, "decode_ms": 13.349708053283393, "receiver_prefill_ms": 170.14595796354115}, "total_wall_ms": 20726.210165885277, "ttft_ms": 13.848332921043038, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ecf5a44f55b5a0c2cfb61d3be37f3d239f5f933dcd1227578cf96f16f839a7be", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "17e1aaeedfb5d9ef640d2d5a", "attempt_id": "0e329e940c1c", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:37:14.677174Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1500}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 1616, "message_tokens": 1524, "message_utf8_bytes": 4770, "cache_tensor_bytes": null, "serialized_payload_bytes": 4770, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 20526.26595797483, "ttft_ms": 13.489958015270531, "decode_ms": 13.593000010587275, "receiver_prefill_ms": 174.63408398907632}, "total_wall_ms": 20730.535249924287, "ttft_ms": 13.489958015270531, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ecf5a44f55b5a0c2cfb61d3be37f3d239f5f933dcd1227578cf96f16f839a7be", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "baadedabdf1638035fbd9759", "attempt_id": "df0ba3b67656", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:37:42.386260Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 29}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 1226, "message_tokens": 1132, "message_utf8_bytes": 2946, "cache_tensor_bytes": null, "serialized_payload_bytes": 2946, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27489.781958051026, "ttft_ms": 13.474542065523565, "decode_ms": 13.51316599175334, "receiver_prefill_ms": 128.8720000302419}, "total_wall_ms": 27647.48529100325, "ttft_ms": 13.474542065523565, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "eb9ca57a5f1773ebf9f8ef83836be66d79f66df18fed066692f9f3ccd4e24a12", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "55a3f4d7fd2c23d9b99b24e6", "attempt_id": "77ea19eee567", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:37:42.559050Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 29}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 1223, "message_tokens": 1132, "message_utf8_bytes": 2946, "cache_tensor_bytes": null, "serialized_payload_bytes": 2946, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27489.781958051026, "ttft_ms": 13.511916040442884, "decode_ms": 13.783792033791542, "receiver_prefill_ms": 139.2971669556573}, "total_wall_ms": 27658.18783303257, "ttft_ms": 13.511916040442884, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "eb9ca57a5f1773ebf9f8ef83836be66d79f66df18fed066692f9f3ccd4e24a12", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "0a032a6aafaafd10ebac965a", "attempt_id": "3ee8388af5bf", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:37:42.741647Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 29}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 1230, "message_tokens": 1132, "message_utf8_bytes": 2946, "cache_tensor_bytes": null, "serialized_payload_bytes": 2946, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27489.781958051026, "ttft_ms": 13.613332994282246, "decode_ms": 120.27341697830707, "receiver_prefill_ms": 483.9315420249477}, "total_wall_ms": 28109.482291038148, "ttft_ms": 13.613332994282246, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "6,914,937", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "eb9ca57a5f1773ebf9f8ef83836be66d79f66df18fed066692f9f3ccd4e24a12", "output_hash": "c56851abfdbf8851226a1e71c4911c616a691c490a883eb754e61bac551d004c"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b7482e845f9639a3ee176238", "attempt_id": "5ba6c7141e87", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:37:43.375433Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 29}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 1223, "message_tokens": 1132, "message_utf8_bytes": 2946, "cache_tensor_bytes": null, "serialized_payload_bytes": 2946, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27489.781958051026, "ttft_ms": 13.337749987840652, "decode_ms": 122.78008298017085, "receiver_prefill_ms": 127.85095896106213}, "total_wall_ms": 27755.644750082865, "ttft_ms": 13.337749987840652, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "6,056,150", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "eb9ca57a5f1773ebf9f8ef83836be66d79f66df18fed066692f9f3ccd4e24a12", "output_hash": "8262889a8478c7e0ee965f159d9849a3849b412e276f5b979476e2da5b320249"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "45f9ff8b9d84bb378b97f725", "attempt_id": "14b307692e32", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:37:43.655827Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 29}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 1233, "message_tokens": 1132, "message_utf8_bytes": 2946, "cache_tensor_bytes": null, "serialized_payload_bytes": 2946, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27489.781958051026, "ttft_ms": 15.502625028602779, "decode_ms": 97.84783306531608, "receiver_prefill_ms": 736.1515830270946}, "total_wall_ms": 28341.198957990855, "ttft_ms": 15.502625028602779, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Lunlith-13", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "eb9ca57a5f1773ebf9f8ef83836be66d79f66df18fed066692f9f3ccd4e24a12", "output_hash": "c45f000afe6d865c8167954eebdff761141225817d4f5b26c962cd50cd75d391"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "9c1193bcd0438a7e0f40d930", "attempt_id": "26e1353a4e6e", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:37:44.522471Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 29}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 1226, "message_tokens": 1132, "message_utf8_bytes": 2946, "cache_tensor_bytes": null, "serialized_payload_bytes": 2946, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27489.781958051026, "ttft_ms": 13.474000035785139, "decode_ms": 93.42749998904765, "receiver_prefill_ms": 128.0549590010196}, "total_wall_ms": 27726.70045797713, "ttft_ms": 13.474000035785139, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1589105", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "eb9ca57a5f1773ebf9f8ef83836be66d79f66df18fed066692f9f3ccd4e24a12", "output_hash": "acce284e21562feb849550dc8d18291445fec57d4ba5820f4463d785f57cda4b"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "7c49e6e0d8d086bfe86ee8cc", "attempt_id": "7290f8250317", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:37:44.776203Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 29}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 1226, "message_tokens": 1132, "message_utf8_bytes": 2946, "cache_tensor_bytes": null, "serialized_payload_bytes": 2946, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27489.781958051026, "ttft_ms": 13.326958054676652, "decode_ms": 13.396458933129907, "receiver_prefill_ms": 127.58637499064207}, "total_wall_ms": 27645.919333095662, "ttft_ms": 13.326958054676652, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "eb9ca57a5f1773ebf9f8ef83836be66d79f66df18fed066692f9f3ccd4e24a12", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "6c9f1e24e67316b8dc41157c", "attempt_id": "548c7468f87f", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:37:44.946758Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 29}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 1228, "message_tokens": 1132, "message_utf8_bytes": 2946, "cache_tensor_bytes": null, "serialized_payload_bytes": 2946, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27489.781958051026, "ttft_ms": 13.738499954342842, "decode_ms": 14.123000088147819, "receiver_prefill_ms": 572.6238330826163}, "total_wall_ms": 28092.158875078894, "ttft_ms": 13.738499954342842, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "eb9ca57a5f1773ebf9f8ef83836be66d79f66df18fed066692f9f3ccd4e24a12", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c4a1a2336088ec54075a5fac", "attempt_id": "331db7ec40b3", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:38:13.185992Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 2576, "message_tokens": 2483, "message_utf8_bytes": 6375, "cache_tensor_bytes": null, "serialized_payload_bytes": 6375, "side_information_bytes": 0, "fetched_bytes": 3418, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27622.584290918894, "ttft_ms": 13.501208974048495, "decode_ms": 13.318917015567422, "receiver_prefill_ms": 343.48633291665465}, "total_wall_ms": 27996.434332919307, "ttft_ms": 13.501208974048495, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "a44d9d91a96f269b4982b6ff2179f485307aba968d8d8b97ed735111b2ae44b0", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "f3b424cffc9c0c35bcb0824b", "attempt_id": "83fd68e12f78", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:38:13.580224Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 2454, "message_tokens": 2364, "message_utf8_bytes": 6181, "cache_tensor_bytes": null, "serialized_payload_bytes": 6181, "side_information_bytes": 0, "fetched_bytes": 3214, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27622.584290918894, "ttft_ms": 14.383958070538938, "decode_ms": 41.993791004642844, "receiver_prefill_ms": 376.38120795600116}, "total_wall_ms": 28058.661248884164, "ttft_ms": 14.383958070538938, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 3}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "DB-7", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "0787dcb36719e1685552d936599bfb3b643e94a5710e7b04a5a35396bdac6a6e", "output_hash": "9197d14a5517d200b55deb6c8141414f24265c8fda55f6ff69a5df1a857e1843"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c08cb1363141a9c90dd4d601", "attempt_id": "9fddbbd0b402", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:38:14.030733Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 2599, "message_tokens": 2502, "message_utf8_bytes": 6413, "cache_tensor_bytes": null, "serialized_payload_bytes": 6413, "side_information_bytes": 0, "fetched_bytes": 3458, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27622.584290918894, "ttft_ms": 13.529125018976629, "decode_ms": 94.11170799285173, "receiver_prefill_ms": 283.0470419721678}, "total_wall_ms": 28016.752248979174, "ttft_ms": 13.529125018976629, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8010205", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "f7192125b3f411c61fac414a6d435b272183a04cae68fc3456b1f0b16a90dacc", "output_hash": "6d1ceb36132f38bd77e8f46c23ea21dd9bcff78bc80bd8a9eaf2d7f2f3bacf46"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b4d1fac0ef1723704ab217b8", "attempt_id": "cc4271d6e3ed", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:38:14.442318Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 2592, "message_tokens": 2502, "message_utf8_bytes": 6413, "cache_tensor_bytes": null, "serialized_payload_bytes": 6413, "side_information_bytes": 0, "fetched_bytes": 3458, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27622.584290918894, "ttft_ms": 14.233208959922194, "decode_ms": 96.33045899681747, "receiver_prefill_ms": 320.92116598505527}, "total_wall_ms": 28057.45874985587, "ttft_ms": 14.233208959922194, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8010205", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "f7192125b3f411c61fac414a6d435b272183a04cae68fc3456b1f0b16a90dacc", "output_hash": "6d1ceb36132f38bd77e8f46c23ea21dd9bcff78bc80bd8a9eaf2d7f2f3bacf46"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "372475fc1e51118002294b99", "attempt_id": "25e0fd2877a3", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:38:14.891428Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 2553, "message_tokens": 2453, "message_utf8_bytes": 6296, "cache_tensor_bytes": null, "serialized_payload_bytes": 6296, "side_information_bytes": 0, "fetched_bytes": 3344, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27622.584290918894, "ttft_ms": 13.588959001936018, "decode_ms": 95.12404189445078, "receiver_prefill_ms": 272.90383295621723}, "total_wall_ms": 28007.517748861574, "ttft_ms": 13.588959001936018, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Lunlith-13", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "e7107c530380b29d0aea1ce99d6f81e40bd82288a5617562f297cf591651cdc2", "output_hash": "c45f000afe6d865c8167954eebdff761141225817d4f5b26c962cd50cd75d391"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "95c3e47e05a1ab9d741f8d76", "attempt_id": "65c3a0cd3984", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:38:15.291134Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 2595, "message_tokens": 2502, "message_utf8_bytes": 6413, "cache_tensor_bytes": null, "serialized_payload_bytes": 6413, "side_information_bytes": 0, "fetched_bytes": 3458, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27622.584290918894, "ttft_ms": 13.745249947533011, "decode_ms": 134.22658306080848, "receiver_prefill_ms": 282.1164169581607}, "total_wall_ms": 28056.00641586352, "ttft_ms": 13.745249947533011, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "21,111,712", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "968b654012b9186fbaa40cd5287808fcfd5329b10f697b795bfcd770a84d5d43", "output_hash": "9d15dcc112d5a8472d06eca5b1d0a107988256fc705559bf1ce1606046ac5e60"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "6106c53fb841f62e4233340b", "attempt_id": "9f06ef650bdd", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:38:15.739420Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 2508, "message_tokens": 2415, "message_utf8_bytes": 6331, "cache_tensor_bytes": null, "serialized_payload_bytes": 6331, "side_information_bytes": 0, "fetched_bytes": 3344, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27622.584290918894, "ttft_ms": 13.826583046466112, "decode_ms": 13.85670795571059, "receiver_prefill_ms": 314.5611659856513}, "total_wall_ms": 27968.10504095629, "ttft_ms": 13.826583046466112, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "4bdbd4670ef7f81c5338ee017970ba7257269aa84389c240aacf8ade520ed9ef", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "08f601663240d6ebff5e8dae", "attempt_id": "88d27028e3fa", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:38:16.099945Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 2578, "message_tokens": 2483, "message_utf8_bytes": 6375, "cache_tensor_bytes": null, "serialized_payload_bytes": 6375, "side_information_bytes": 0, "fetched_bytes": 3418, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27622.584290918894, "ttft_ms": 13.81716609466821, "decode_ms": 13.62275006249547, "receiver_prefill_ms": 291.97179106995463}, "total_wall_ms": 27945.460707880557, "ttft_ms": 13.81716609466821, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "c379375be543f7921a862922d44f13ca776a3adc25a8fb4dbf59f298c01e9a31", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "91ed6e8a7351d6c3f936996a", "attempt_id": "50d72987c791", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:38:16.840375Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 95404032}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 8738, "cache_tensor_bytes": 95404032, "serialized_payload_bytes": 95404032, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 402.01320801861584, "ttft_ms": 33.228709013201296, "decode_ms": 84.95041693095118, "receiver_prefill_ms": 0.0}, "total_wall_ms": 523.9856659900397, "ttft_ms": 33.228709013201296, "input_usage": {}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "535300", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "fc689fe95e517c971459f8d7a6abbbf8873a56ea6a4493a05c86e332dc7659e2"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "5900a046d7d45e371663ec75", "attempt_id": "6df4c1e22108", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:38:16.977315Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 95404032}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 44, "message_tokens": null, "message_utf8_bytes": 8738, "cache_tensor_bytes": 95404032, "serialized_payload_bytes": 95404032, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 402.01320801861584, "ttft_ms": 28.73795898631215, "decode_ms": 139.19545896351337, "receiver_prefill_ms": 0.0}, "total_wall_ms": 573.0392909608781, "ttft_ms": 28.73795898631215, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "DD-69N-1957", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "c7a79bc9b8b76dec6be02055bce804a678265ed054fadbc587c4eb56d55e06dd"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "20502c201a6d13ba456edc83", "attempt_id": "b22bfd06811f", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:38:17.163210Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 95404032}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 51, "message_tokens": null, "message_utf8_bytes": 8738, "cache_tensor_bytes": 95404032, "serialized_payload_bytes": 95404032, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 402.01320801861584, "ttft_ms": 30.972167034633458, "decode_ms": 101.72241600230336, "receiver_prefill_ms": 0.0}, "total_wall_ms": 537.7437500283122, "ttft_ms": 30.972167034633458, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5632215", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "dff2a8edf1f157947cce1a41ebc57f74d76214f6c6a3ee4b8447fd1333253dba"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "bc18051a4cbe7f14bd4ad2a7", "attempt_id": "78031e1fdbbf", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:38:17.313415Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 95404032}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 44, "message_tokens": null, "message_utf8_bytes": 8738, "cache_tensor_bytes": 95404032, "serialized_payload_bytes": 95404032, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 402.01320801861584, "ttft_ms": 28.50220794789493, "decode_ms": 97.76454197708517, "receiver_prefill_ms": 0.0}, "total_wall_ms": 531.184250023216, "ttft_ms": 28.50220794789493, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5632215", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "dff2a8edf1f157947cce1a41ebc57f74d76214f6c6a3ee4b8447fd1333253dba"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "906b85ba834d280cb5fa428f", "attempt_id": "9bdb715d3bac", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:38:17.457854Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 95404032}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 54, "message_tokens": null, "message_utf8_bytes": 8738, "cache_tensor_bytes": 95404032, "serialized_payload_bytes": 95404032, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 402.01320801861584, "ttft_ms": 29.76112498436123, "decode_ms": 97.67175000160933, "receiver_prefill_ms": 0.0}, "total_wall_ms": 532.5968739343807, "ttft_ms": 29.76112498436123, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Lunlith-13", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "c45f000afe6d865c8167954eebdff761141225817d4f5b26c962cd50cd75d391"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "3e334213ca93c286c2f6f01b", "attempt_id": "1db7478084a7", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:38:17.603010Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 95404032}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 8738, "cache_tensor_bytes": 95404032, "serialized_payload_bytes": 95404032, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 402.01320801861584, "ttft_ms": 29.825582983903587, "decode_ms": 111.74766696058214, "receiver_prefill_ms": 0.0}, "total_wall_ms": 546.6306660091504, "ttft_ms": 29.825582983903587, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "14212622", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "dedcb11be6d709824001d75c7e619b4850c7ba6f3d8f91811d8586f369d66903"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "7ffa379a6e82d2a4053549f0", "attempt_id": "7d577407b505", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:38:17.762321Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 95404032}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 8738, "cache_tensor_bytes": 95404032, "serialized_payload_bytes": 95404032, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 402.01320801861584, "ttft_ms": 29.81837501283735, "decode_ms": 13.92462500371039, "receiver_prefill_ms": 0.0}, "total_wall_ms": 449.0244999760762, "ttft_ms": 29.81837501283735, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "de91d30798274bed9ddca66d", "attempt_id": "7bf8b0911391", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:38:17.823560Z", "task_family": "synthetic", "document_id": "syn_03027_2000", "document_hash": "10326080979f71b625e4110c7c19767e5a752e267de5034f7e37d98cb6e06249", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03027_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 95404032}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1955, "query_tokens": 49, "message_tokens": null, "message_utf8_bytes": 8738, "cache_tensor_bytes": 95404032, "serialized_payload_bytes": 95404032, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 402.01320801861584, "ttft_ms": 30.746832955628633, "decode_ms": 13.972665998153389, "receiver_prefill_ms": 0.0}, "total_wall_ms": 449.8988330597058, "ttft_ms": 30.746832955628633, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Sollith-4", "bogus_amount": "5,632,215", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03027_2000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "cd1de282312e9a6a4a36da26", "attempt_id": "95eb39e50ada", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:38:18.818081Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 6249, "message_tokens": 6163, "message_utf8_bytes": 30147, "cache_tensor_bytes": null, "serialized_payload_bytes": 30147, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.035415985621511936, "ttft_ms": 13.989083003252745, "decode_ms": 109.78400008752942, "receiver_prefill_ms": 744.913375005126}, "total_wall_ms": 878.7341250572354, "ttft_ms": 13.989083003252745, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "549,950 EUR", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 0.6666666666666666}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "900139fea49a30dd0078ecbe42e8fa0dbddfd1684d4c3fe6d000ec06147f0ac3", "output_hash": "c4834b6c2ddcc337bd255fb6f2c9f51520d4e94c30b827d39f041d1e90d764ce"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "295ec1ee200cf671905ded5b", "attempt_id": "0c92bb039dc2", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:38:19.713311Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 6250, "message_tokens": 6163, "message_utf8_bytes": 30147, "cache_tensor_bytes": null, "serialized_payload_bytes": 30147, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.035415985621511936, "ttft_ms": 14.207542059011757, "decode_ms": 138.16983392462134, "receiver_prefill_ms": 731.6718340152875}, "total_wall_ms": 893.9878739183769, "ttft_ms": 14.207542059011757, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "PB-38S-9552", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "900139fea49a30dd0078ecbe42e8fa0dbddfd1684d4c3fe6d000ec06147f0ac3", "output_hash": "2ead15278e346194e16500a37d0e52a9739df0cf36c158fed043974d95161e89"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c68e6d3679a36604723b6a96", "attempt_id": "3dc3fef7e3ed", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:38:20.624397Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 6257, "message_tokens": 6163, "message_utf8_bytes": 30147, "cache_tensor_bytes": null, "serialized_payload_bytes": 30147, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.035415985621511936, "ttft_ms": 13.839625054970384, "decode_ms": 95.94604093581438, "receiver_prefill_ms": 732.6376250712201}, "total_wall_ms": 851.8903750227764, "ttft_ms": 13.839625054970384, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "497,425", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "900139fea49a30dd0078ecbe42e8fa0dbddfd1684d4c3fe6d000ec06147f0ac3", "output_hash": "88a6c470039e48606cf3771bfcac5aa93ce7a2bdb9bded7ea7861cbc2208fe4a"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "de386e89f96871a8d0e8036e", "attempt_id": "46ff4a14b8d2", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:38:21.493535Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 6250, "message_tokens": 6163, "message_utf8_bytes": 30147, "cache_tensor_bytes": null, "serialized_payload_bytes": 30147, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.035415985621511936, "ttft_ms": 13.864375068806112, "decode_ms": 96.03904106188565, "receiver_prefill_ms": 732.1783329825848}, "total_wall_ms": 851.7837909748778, "ttft_ms": 13.864375068806112, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "497,425", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "900139fea49a30dd0078ecbe42e8fa0dbddfd1684d4c3fe6d000ec06147f0ac3", "output_hash": "88a6c470039e48606cf3771bfcac5aa93ce7a2bdb9bded7ea7861cbc2208fe4a"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a3f7ad015a81b5005bc9410d", "attempt_id": "cbe9e00d5e26", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:38:22.361877Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 6257, "message_tokens": 6163, "message_utf8_bytes": 30147, "cache_tensor_bytes": null, "serialized_payload_bytes": 30147, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.035415985621511936, "ttft_ms": 14.23920807428658, "decode_ms": 70.00312500167638, "receiver_prefill_ms": 735.4718339629471}, "total_wall_ms": 829.2623319430277, "ttft_ms": 14.23920807428658, "input_usage": {}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Felard-13", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "900139fea49a30dd0078ecbe42e8fa0dbddfd1684d4c3fe6d000ec06147f0ac3", "output_hash": "3148a4375aedadf07a17298878dde115102f658fb7bebe944a013c9b53bec240"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "91871f11ab4959fff81aad87", "attempt_id": "af4e64001216", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:38:23.207711Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 6252, "message_tokens": 6163, "message_utf8_bytes": 30147, "cache_tensor_bytes": null, "serialized_payload_bytes": 30147, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.035415985621511936, "ttft_ms": 13.687415979802608, "decode_ms": 138.01412505563349, "receiver_prefill_ms": 742.2421249793842}, "total_wall_ms": 903.5595409804955, "ttft_ms": 13.687415979802608, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "44,149,995", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "900139fea49a30dd0078ecbe42e8fa0dbddfd1684d4c3fe6d000ec06147f0ac3", "output_hash": "40a1919ac43e538e1d9037d7b31dc9a2d788e76c77cf70f89b16b2b6ad6f7b18"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "89489c4a88516b8a8acc4062", "attempt_id": "ad001c067984", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:38:24.128524Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 6252, "message_tokens": 6163, "message_utf8_bytes": 30147, "cache_tensor_bytes": null, "serialized_payload_bytes": 30147, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.035415985621511936, "ttft_ms": 13.56999995186925, "decode_ms": 13.789708027616143, "receiver_prefill_ms": 740.81574997399}, "total_wall_ms": 777.688249014318, "ttft_ms": 13.56999995186925, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "900139fea49a30dd0078ecbe42e8fa0dbddfd1684d4c3fe6d000ec06147f0ac3", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "f6dcbb853ee3aef0d49e4953", "attempt_id": "eced7537b00d", "run_group_id": "E5_core:full_context|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:38:24.923140Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 6254, "message_tokens": 6163, "message_utf8_bytes": 30147, "cache_tensor_bytes": null, "serialized_payload_bytes": 30147, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.035415985621511936, "ttft_ms": 13.857875019311905, "decode_ms": 13.719791080802679, "receiver_prefill_ms": 748.0398750631139}, "total_wall_ms": 785.2222908986732, "ttft_ms": 13.857875019311905, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "900139fea49a30dd0078ecbe42e8fa0dbddfd1684d4c3fe6d000ec06147f0ac3", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "985d0ab443c33b552f4d7bb1", "attempt_id": "7455d29f9fd2", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:38:40.291683Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 1124, "message_tokens": 1037, "message_utf8_bytes": 2825, "cache_tensor_bytes": null, "serialized_payload_bytes": 2825, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14565.942666959018, "ttft_ms": 13.322041020728648, "decode_ms": 94.27224996034056, "receiver_prefill_ms": 121.7627499718219}, "total_wall_ms": 14797.277083969675, "ttft_ms": 13.322041020728648, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "497,425", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "151d065dca4aa59d8fb9d6e010f80e0737ece0c9b7924ad35c32e329f35af84d", "output_hash": "88a6c470039e48606cf3771bfcac5aa93ce7a2bdb9bded7ea7861cbc2208fe4a"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c4ef7cd9ac11f9f4e66264b3", "attempt_id": "b8471fd35a4e", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:38:40.540646Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 1125, "message_tokens": 1037, "message_utf8_bytes": 2825, "cache_tensor_bytes": null, "serialized_payload_bytes": 2825, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14565.942666959018, "ttft_ms": 13.746540993452072, "decode_ms": 135.76487498357892, "receiver_prefill_ms": 120.58391701430082}, "total_wall_ms": 14838.071958976798, "ttft_ms": 13.746540993452072, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "PB-38S-9552", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "151d065dca4aa59d8fb9d6e010f80e0737ece0c9b7924ad35c32e329f35af84d", "output_hash": "2ead15278e346194e16500a37d0e52a9739df0cf36c158fed043974d95161e89"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b626bbd9643b5f405dbef136", "attempt_id": "a569463fbd49", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:38:40.830438Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 1132, "message_tokens": 1037, "message_utf8_bytes": 2825, "cache_tensor_bytes": null, "serialized_payload_bytes": 2825, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14565.942666959018, "ttft_ms": 13.365166960284114, "decode_ms": 94.74395797587931, "receiver_prefill_ms": 121.7123749665916}, "total_wall_ms": 14797.758791944943, "ttft_ms": 13.365166960284114, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "603,595", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "151d065dca4aa59d8fb9d6e010f80e0737ece0c9b7924ad35c32e329f35af84d", "output_hash": "19f63ee7169f83e8e0256e7a4a4a68fa5980ea55ef086b8b373177c1d02539bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "d4124e6075f8a9604c646581", "attempt_id": "ff011d3d49ef", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:38:41.083278Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 1125, "message_tokens": 1037, "message_utf8_bytes": 2825, "cache_tensor_bytes": null, "serialized_payload_bytes": 2825, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14565.942666959018, "ttft_ms": 13.358833035454154, "decode_ms": 94.34662503190339, "receiver_prefill_ms": 120.74537493754178}, "total_wall_ms": 14796.287126024254, "ttft_ms": 13.358833035454154, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "497,425", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "151d065dca4aa59d8fb9d6e010f80e0737ece0c9b7924ad35c32e329f35af84d", "output_hash": "88a6c470039e48606cf3771bfcac5aa93ce7a2bdb9bded7ea7861cbc2208fe4a"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "35dff6aa97f8d092dba3540e", "attempt_id": "1c8c7bab14f5", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:38:41.329957Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 1132, "message_tokens": 1037, "message_utf8_bytes": 2825, "cache_tensor_bytes": null, "serialized_payload_bytes": 2825, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14565.942666959018, "ttft_ms": 13.397749979048967, "decode_ms": 67.36020802054554, "receiver_prefill_ms": 120.75937504414469}, "total_wall_ms": 14769.339166930877, "ttft_ms": 13.397749979048967, "input_usage": {}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Felard-13", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "151d065dca4aa59d8fb9d6e010f80e0737ece0c9b7924ad35c32e329f35af84d", "output_hash": "3148a4375aedadf07a17298878dde115102f658fb7bebe944a013c9b53bec240"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a43d83eb7cb29f177523b020", "attempt_id": "2bbce92abfc6", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:38:41.550373Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 1127, "message_tokens": 1037, "message_utf8_bytes": 2825, "cache_tensor_bytes": null, "serialized_payload_bytes": 2825, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14565.942666959018, "ttft_ms": 13.514458085410297, "decode_ms": 120.91212498489767, "receiver_prefill_ms": 121.30154203623533}, "total_wall_ms": 14823.65087501239, "ttft_ms": 13.514458085410297, "input_usage": {}, "output_usage": {"output_tokens": 9}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "34024.145", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "151d065dca4aa59d8fb9d6e010f80e0737ece0c9b7924ad35c32e329f35af84d", "output_hash": "2b4cea98a1c37694861ded9200fa7c12e5226877d9b114d124bd5aac8270b076"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "d7c846df87fe7db476adb024", "attempt_id": "aab2180fe453", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:38:41.825295Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 1127, "message_tokens": 1037, "message_utf8_bytes": 2825, "cache_tensor_bytes": null, "serialized_payload_bytes": 2825, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14565.942666959018, "ttft_ms": 13.363166013732553, "decode_ms": 13.457249966450036, "receiver_prefill_ms": 121.40012509189546}, "total_wall_ms": 14716.055458877236, "ttft_ms": 13.363166013732553, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "151d065dca4aa59d8fb9d6e010f80e0737ece0c9b7924ad35c32e329f35af84d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "0c45664e9fbf127d3bf8c0d6", "attempt_id": "a8509fc5b506", "run_group_id": "E5_core:concat_A4|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:38:41.992866Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "sender_tokens": 1024}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 1129, "message_tokens": 1037, "message_utf8_bytes": 2825, "cache_tensor_bytes": null, "serialized_payload_bytes": 2825, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 14565.942666959018, "ttft_ms": 13.547584065236151, "decode_ms": 13.681709067896008, "receiver_prefill_ms": 121.35795806534588}, "total_wall_ms": 14716.467458987609, "ttft_ms": 13.547584065236151, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "151d065dca4aa59d8fb9d6e010f80e0737ece0c9b7924ad35c32e329f35af84d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "112a07bbaa686c178d2f7524", "attempt_id": "e76da0d9998c", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:38:54.215525Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 11}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 535, "message_tokens": 444, "message_utf8_bytes": 1193, "cache_tensor_bytes": null, "serialized_payload_bytes": 1193, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 12055.380290956236, "ttft_ms": 13.085917104035616, "decode_ms": 13.471207930706441, "receiver_prefill_ms": 64.34062495827675}, "total_wall_ms": 12153.142706956714, "ttft_ms": 13.085917104035616, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "51175a30c775ae640d138df42dd84cc82b8f6f0eef31d48ae508ea70fd116fe5", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "9f32b62d17eb7c03439e5dbb", "attempt_id": "11fd468c0407", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:38:54.330882Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 11}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 536, "message_tokens": 444, "message_utf8_bytes": 1193, "cache_tensor_bytes": null, "serialized_payload_bytes": 1193, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 12055.380290956236, "ttft_ms": 13.279999955557287, "decode_ms": 13.302500010468066, "receiver_prefill_ms": 64.39266703091562}, "total_wall_ms": 12147.493958007544, "ttft_ms": 13.279999955557287, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "51175a30c775ae640d138df42dd84cc82b8f6f0eef31d48ae508ea70fd116fe5", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a9d42622602eed42ca94a8bc", "attempt_id": "c7de5b15b398", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:38:54.440591Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 11}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 543, "message_tokens": 444, "message_utf8_bytes": 1193, "cache_tensor_bytes": null, "serialized_payload_bytes": 1193, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 12055.380290956236, "ttft_ms": 13.592165894806385, "decode_ms": 13.167624943889678, "receiver_prefill_ms": 64.24829200841486}, "total_wall_ms": 12147.520790924318, "ttft_ms": 13.592165894806385, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "51175a30c775ae640d138df42dd84cc82b8f6f0eef31d48ae508ea70fd116fe5", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "288716e136cfedb090e0c4d8", "attempt_id": "21709a393c98", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:38:54.550311Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 11}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 536, "message_tokens": 444, "message_utf8_bytes": 1193, "cache_tensor_bytes": null, "serialized_payload_bytes": 1193, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 12055.380290956236, "ttft_ms": 13.230041950009763, "decode_ms": 13.309958972968161, "receiver_prefill_ms": 64.05245803762227}, "total_wall_ms": 12147.104832925834, "ttft_ms": 13.230041950009763, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "51175a30c775ae640d138df42dd84cc82b8f6f0eef31d48ae508ea70fd116fe5", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "1860a27b19ed37d0444448ba", "attempt_id": "8d8ab541c15c", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:38:54.659154Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 11}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 543, "message_tokens": 444, "message_utf8_bytes": 1193, "cache_tensor_bytes": null, "serialized_payload_bytes": 1193, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 12055.380290956236, "ttft_ms": 13.31208401825279, "decode_ms": 67.04100000206381, "receiver_prefill_ms": 64.25454199779779}, "total_wall_ms": 12201.149957953021, "ttft_ms": 13.31208401825279, "input_usage": {}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Felard-13", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "51175a30c775ae640d138df42dd84cc82b8f6f0eef31d48ae508ea70fd116fe5", "output_hash": "3148a4375aedadf07a17298878dde115102f658fb7bebe944a013c9b53bec240"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "fb31ea512ba7e3ed7860d93c", "attempt_id": "8d73212957d8", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:38:54.821917Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 11}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 538, "message_tokens": 444, "message_utf8_bytes": 1193, "cache_tensor_bytes": null, "serialized_payload_bytes": 1193, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 12055.380290956236, "ttft_ms": 13.527916045859456, "decode_ms": 93.64537498913705, "receiver_prefill_ms": 64.35816595330834}, "total_wall_ms": 12228.103874949738, "ttft_ms": 13.527916045859456, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "2333595", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "51175a30c775ae640d138df42dd84cc82b8f6f0eef31d48ae508ea70fd116fe5", "output_hash": "e1a46be96c350176b34fe6c169a1d48a9f809d6e87325d053a5955ec1485f300"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "8f61d2cdf9219975703a1287", "attempt_id": "a28474b40af4", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:38:55.011932Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 11}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 538, "message_tokens": 444, "message_utf8_bytes": 1193, "cache_tensor_bytes": null, "serialized_payload_bytes": 1193, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 12055.380290956236, "ttft_ms": 13.129665981978178, "decode_ms": 13.177125016227365, "receiver_prefill_ms": 64.38358302693814}, "total_wall_ms": 12147.220165934414, "ttft_ms": 13.129665981978178, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "51175a30c775ae640d138df42dd84cc82b8f6f0eef31d48ae508ea70fd116fe5", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "9090c8fcd986e8b59e78c9d2", "attempt_id": "98be251ebda7", "run_group_id": "E5_core:merge_A5|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:38:55.121224Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512, "n_facts": 11}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 540, "message_tokens": 444, "message_utf8_bytes": 1193, "cache_tensor_bytes": null, "serialized_payload_bytes": 1193, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 12055.380290956236, "ttft_ms": 13.198708067648113, "decode_ms": 13.752583065070212, "receiver_prefill_ms": 64.11479099188}, "total_wall_ms": 12147.59041601792, "ttft_ms": 13.198708067648113, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "51175a30c775ae640d138df42dd84cc82b8f6f0eef31d48ae508ea70fd116fe5", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "93e244f6fde9f68053d96564", "attempt_id": "0070e40ff40e", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:39:07.306815Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 1319, "message_tokens": 1229, "message_utf8_bytes": 3083, "cache_tensor_bytes": null, "serialized_payload_bytes": 3083, "side_information_bytes": 0, "fetched_bytes": 1729, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 12073.418041924015, "ttft_ms": 13.415541965514421, "decode_ms": 94.07912497408688, "receiver_prefill_ms": 143.16058298572898}, "total_wall_ms": 12329.695999971591, "ttft_ms": 13.415541965514421, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "549.950", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "c32204796940cd9059911b0e83a82fefb88f14e6e434a96164278cc3586a0353", "output_hash": "cab784b9a3b9aa92b5218ac427b87c725bb99e22c3af34061c9cf715ac2eedec"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "e2cbea7f0cc29b66c0aaa799", "attempt_id": "5dbedd40409e", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:39:07.581130Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 1304, "message_tokens": 1213, "message_utf8_bytes": 3040, "cache_tensor_bytes": null, "serialized_payload_bytes": 3040, "side_information_bytes": 0, "fetched_bytes": 1690, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 12073.418041924015, "ttft_ms": 13.533500023186207, "decode_ms": 54.50850003398955, "receiver_prefill_ms": 140.57745796162635}, "total_wall_ms": 12284.050791990012, "ttft_ms": 13.533500023186207, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 4}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "LN-61", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "f5d3ba2174a7933770db075e755db772ac6db07dea7ffad7f17132b91092c6ad", "output_hash": "bd0c9af579a47956226f82ea004c3a4d7a6bcfa10004153a5143219fb282c26c"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "f4a33279ca4eb17ee6d5da29", "attempt_id": "17d120cad191", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:39:07.811498Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 1311, "message_tokens": 1213, "message_utf8_bytes": 3040, "cache_tensor_bytes": null, "serialized_payload_bytes": 3040, "side_information_bytes": 0, "fetched_bytes": 1690, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 12073.418041924015, "ttft_ms": 13.467582990415394, "decode_ms": 80.98425006028265, "receiver_prefill_ms": 140.87254100013524}, "total_wall_ms": 12310.869166860357, "ttft_ms": 13.467582990415394, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "497425", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ad8be07e5b810377e6800b6c9786cc64b635c6a10a5541027f66ffcb67b88b41", "output_hash": "12747c557f66bc4ad85f9222ae86718a2c1a8fcfd94c6943ecda0388a8a62473"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "27680b92686267fd08732c9c", "attempt_id": "11d9290972a9", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:39:08.066033Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 1304, "message_tokens": 1213, "message_utf8_bytes": 3040, "cache_tensor_bytes": null, "serialized_payload_bytes": 3040, "side_information_bytes": 0, "fetched_bytes": 1690, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 12073.418041924015, "ttft_ms": 13.292624964378774, "decode_ms": 95.05629097111523, "receiver_prefill_ms": 140.8732089912519}, "total_wall_ms": 12324.647166882642, "ttft_ms": 13.292624964378774, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "497,425", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "ad8be07e5b810377e6800b6c9786cc64b635c6a10a5541027f66ffcb67b88b41", "output_hash": "88a6c470039e48606cf3771bfcac5aa93ce7a2bdb9bded7ea7861cbc2208fe4a"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "950b6b0c843d609b4b1bb3eb", "attempt_id": "6b4194376f6e", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:39:08.334905Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 1301, "message_tokens": 1203, "message_utf8_bytes": 3041, "cache_tensor_bytes": null, "serialized_payload_bytes": 3041, "side_information_bytes": 0, "fetched_bytes": 1694, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 12073.418041924015, "ttft_ms": 14.588250080123544, "decode_ms": 67.64995900448412, "receiver_prefill_ms": 140.38187498226762}, "total_wall_ms": 12298.478957964107, "ttft_ms": 14.588250080123544, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Felard-13", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "806f885c7de2a1808b8f0b040599be99eae323449060b87b742b977ff684ac83", "output_hash": "3148a4375aedadf07a17298878dde115102f658fb7bebe944a013c9b53bec240"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "edc0b1be7c9f7fa7c2c7e20d", "attempt_id": "e9b805025549", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:39:08.577401Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 1322, "message_tokens": 1229, "message_utf8_bytes": 3083, "cache_tensor_bytes": null, "serialized_payload_bytes": 3083, "side_information_bytes": 0, "fetched_bytes": 1729, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 12073.418041924015, "ttft_ms": 13.765249983407557, "decode_ms": 134.42337501328439, "receiver_prefill_ms": 143.2016659528017}, "total_wall_ms": 12366.888041840866, "ttft_ms": 13.765249983407557, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "33,378,635", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "c32204796940cd9059911b0e83a82fefb88f14e6e434a96164278cc3586a0353", "output_hash": "efb67ab407abdf4f461142e19cb21cb95eaf36d78f7ea473dcc3a5e6dcfc912e"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "6bac3997fe1f705049e79dbb", "attempt_id": "0ae48d7b5fcd", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:39:08.888223Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 1306, "message_tokens": 1213, "message_utf8_bytes": 3040, "cache_tensor_bytes": null, "serialized_payload_bytes": 3040, "side_information_bytes": 0, "fetched_bytes": 1690, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 12073.418041924015, "ttft_ms": 13.826625072397292, "decode_ms": 13.40545795392245, "receiver_prefill_ms": 140.46274998690933}, "total_wall_ms": 12243.150208843872, "ttft_ms": 13.826625072397292, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "f5d3ba2174a7933770db075e755db772ac6db07dea7ffad7f17132b91092c6ad", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "0ea656f271616a1bdcdaf0ab", "attempt_id": "c332b4b86be9", "run_group_id": "E5_core:shared_memory_A6|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:39:09.075180Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 2, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 1308, "message_tokens": 1213, "message_utf8_bytes": 3040, "cache_tensor_bytes": null, "serialized_payload_bytes": 3040, "side_information_bytes": 0, "fetched_bytes": 1690, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 12073.418041924015, "ttft_ms": 13.667916995473206, "decode_ms": 13.608917011879385, "receiver_prefill_ms": 140.2309580007568}, "total_wall_ms": 12242.982750874944, "ttft_ms": 13.667916995473206, "input_usage": {"lookups": 6}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "f5d3ba2174a7933770db075e755db772ac6db07dea7ffad7f17132b91092c6ad", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "406e868916da7e7f0b8e1196", "attempt_id": "0f0b258d752b", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:39:10.002086Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 230068224}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 44, "message_tokens": null, "message_utf8_bytes": 30121, "cache_tensor_bytes": 230068224, "serialized_payload_bytes": 230068224, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 739.9438340216875, "ttft_ms": 53.17895789630711, "decode_ms": 108.01137494854629, "receiver_prefill_ms": 0.0}, "total_wall_ms": 910.6197509681806, "ttft_ms": 53.17895789630711, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "549,950", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "c186ae04f377e0a3f95a6b3affbd1003677f14085f9e34f17926f3b1d1a83d6a"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "5141a04d9dfff8ac67ae46e4", "attempt_id": "07548b8fdef8", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:39:10.190143Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 230068224}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 45, "message_tokens": null, "message_utf8_bytes": 30121, "cache_tensor_bytes": 230068224, "serialized_payload_bytes": 230068224, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 739.9438340216875, "ttft_ms": 47.442832961678505, "decode_ms": 150.01800004392862, "receiver_prefill_ms": 0.0}, "total_wall_ms": 944.9041261104867, "ttft_ms": 47.442832961678505, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "PB-38S-9552", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "2ead15278e346194e16500a37d0e52a9739df0cf36c158fed043974d95161e89"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "41f77c720100b784ca72c645", "attempt_id": "5a3a4cb9347c", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:39:10.412303Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 230068224}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 52, "message_tokens": null, "message_utf8_bytes": 30121, "cache_tensor_bytes": 230068224, "serialized_payload_bytes": 230068224, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 739.9438340216875, "ttft_ms": 47.251207986846566, "decode_ms": 120.6780830398202, "receiver_prefill_ms": 0.0}, "total_wall_ms": 915.571584017016, "ttft_ms": 47.251207986846566, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "10066730", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "87519a500211c8cb5f53d6b4cda2cd518b18d757f7adef0e41cb4fe4f1ce9bb3"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "fe9d5a5107112d50cbf8f6e6", "attempt_id": "90cc0af23d0e", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:39:10.605113Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 230068224}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 45, "message_tokens": null, "message_utf8_bytes": 30121, "cache_tensor_bytes": 230068224, "serialized_payload_bytes": 230068224, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 739.9438340216875, "ttft_ms": 46.87399999238551, "decode_ms": 104.96341693215072, "receiver_prefill_ms": 0.0}, "total_wall_ms": 899.2280840175226, "ttft_ms": 46.87399999238551, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "497,425", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "88a6c470039e48606cf3771bfcac5aa93ce7a2bdb9bded7ea7861cbc2208fe4a"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "d7bbe6e2f75a6328aa2c0293", "attempt_id": "b6c5e4040fe6", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:39:10.782210Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 230068224}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 52, "message_tokens": null, "message_utf8_bytes": 30121, "cache_tensor_bytes": 230068224, "serialized_payload_bytes": 230068224, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 739.9438340216875, "ttft_ms": 46.924500027671456, "decode_ms": 74.9058750225231, "receiver_prefill_ms": 0.0}, "total_wall_ms": 869.6087090065703, "ttft_ms": 46.924500027671456, "input_usage": {}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Felard-13", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "3148a4375aedadf07a17298878dde115102f658fb7bebe944a013c9b53bec240"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "33493e99c11456e6b8bbbe02", "attempt_id": "2c7b13c588f2", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:39:10.929357Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 230068224}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 30121, "cache_tensor_bytes": 230068224, "serialized_payload_bytes": 230068224, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 739.9438340216875, "ttft_ms": 45.90945807285607, "decode_ms": 119.92999992799014, "receiver_prefill_ms": 0.0}, "total_wall_ms": 913.1274591200054, "ttft_ms": 45.90945807285607, "input_usage": {}, "output_usage": {"output_tokens": 8}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "14639955", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "c5633f58fba4e80c2230407774334bab5798fc116be8eb85f77902cff7f00c0c"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "cb98c89526a7c964c2529c2b", "attempt_id": "6529dd68d0ac", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:39:11.120163Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 230068224}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 30121, "cache_tensor_bytes": 230068224, "serialized_payload_bytes": 230068224, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 739.9438340216875, "ttft_ms": 45.78416596632451, "decode_ms": 14.98570793773979, "receiver_prefill_ms": 0.0}, "total_wall_ms": 808.1674589775503, "ttft_ms": 45.78416596632451, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "18c2fd6f3252f316fbe7ff35", "attempt_id": "62af46506132", "run_group_id": "E5_core:cache_concat_negative_control|k=2", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:39:11.205530Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 2, "budget_tokens": 512, "cache_bytes": 230068224}, "question_access": "blind", "topology": "fanin_k2", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 49, "message_tokens": null, "message_utf8_bytes": 30121, "cache_tensor_bytes": 230068224, "serialized_payload_bytes": 230068224, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 739.9438340216875, "ttft_ms": 47.30895906686783, "decode_ms": 15.242750057950616, "receiver_prefill_ms": 0.0}, "total_wall_ms": 810.0527919596061, "ttft_ms": 47.30895906686783, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "68e16b868a8fca5c98f45431", "attempt_id": "eca64d427607", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:39:11.292807Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 6363, "message_tokens": 6277, "message_utf8_bytes": 30651, "cache_tensor_bytes": null, "serialized_payload_bytes": 30651, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.0314170029014349, "ttft_ms": 13.798166997730732, "decode_ms": 97.40820899605751, "receiver_prefill_ms": 742.2527919989079}, "total_wall_ms": 863.2012089947239, "ttft_ms": 13.798166997730732, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "549,950", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "9395f0e47fc863280e77b9f57d6ff310ff74ff0f4a397f483b809756f2aee3a2", "output_hash": "c186ae04f377e0a3f95a6b3affbd1003677f14085f9e34f17926f3b1d1a83d6a"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c8afec8f24ecc90d4a5b54e2", "attempt_id": "2c18ed011945", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:39:12.173004Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 6364, "message_tokens": 6277, "message_utf8_bytes": 30651, "cache_tensor_bytes": null, "serialized_payload_bytes": 30651, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.0314170029014349, "ttft_ms": 13.695958070456982, "decode_ms": 138.09312495868653, "receiver_prefill_ms": 743.1468750583008}, "total_wall_ms": 904.6748749678954, "ttft_ms": 13.695958070456982, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "PB-38S-9552", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "9395f0e47fc863280e77b9f57d6ff310ff74ff0f4a397f483b809756f2aee3a2", "output_hash": "2ead15278e346194e16500a37d0e52a9739df0cf36c158fed043974d95161e89"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a1abaa66c9c3fe53954cd2da", "attempt_id": "bd33d2657af7", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:39:13.094676Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 6371, "message_tokens": 6277, "message_utf8_bytes": 30651, "cache_tensor_bytes": null, "serialized_payload_bytes": 30651, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.0314170029014349, "ttft_ms": 14.193541021086276, "decode_ms": 96.49883303791285, "receiver_prefill_ms": 758.2171249669045}, "total_wall_ms": 878.7434579571709, "ttft_ms": 14.193541021086276, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "497,425", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "9395f0e47fc863280e77b9f57d6ff310ff74ff0f4a397f483b809756f2aee3a2", "output_hash": "88a6c470039e48606cf3771bfcac5aa93ce7a2bdb9bded7ea7861cbc2208fe4a"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "f0d2cd8c7fa52d0cdae8528e", "attempt_id": "c0ce2ad94d05", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:39:13.990550Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 6364, "message_tokens": 6277, "message_utf8_bytes": 30651, "cache_tensor_bytes": null, "serialized_payload_bytes": 30651, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.0314170029014349, "ttft_ms": 13.974832952953875, "decode_ms": 96.81954199913889, "receiver_prefill_ms": 758.3319169934839}, "total_wall_ms": 878.825583960861, "ttft_ms": 13.974832952953875, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "497,425", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "9395f0e47fc863280e77b9f57d6ff310ff74ff0f4a397f483b809756f2aee3a2", "output_hash": "88a6c470039e48606cf3771bfcac5aa93ce7a2bdb9bded7ea7861cbc2208fe4a"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c411f61cbef92836fd2fb62e", "attempt_id": "c125c7f7aa0d", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:39:14.886531Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 6371, "message_tokens": 6277, "message_utf8_bytes": 30651, "cache_tensor_bytes": null, "serialized_payload_bytes": 30651, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.0314170029014349, "ttft_ms": 14.021249953657389, "decode_ms": 56.06083304155618, "receiver_prefill_ms": 767.2095829620957}, "total_wall_ms": 846.9825000502169, "ttft_ms": 14.021249953657389, "input_usage": {}, "output_usage": {"output_tokens": 4}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Melqui-6", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "9395f0e47fc863280e77b9f57d6ff310ff74ff0f4a397f483b809756f2aee3a2", "output_hash": "61840358f56e26ce55e72905cc4adc8d6434d4ed2c50a6b7dc75ccb1d7f170a7"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "db77e13438d75ab69e60fe7e", "attempt_id": "efcebbceeebb", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:39:15.750524Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 6366, "message_tokens": 6277, "message_utf8_bytes": 30651, "cache_tensor_bytes": null, "serialized_payload_bytes": 30651, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.0314170029014349, "ttft_ms": 14.07766598276794, "decode_ms": 96.71724995132536, "receiver_prefill_ms": 778.0761249596253}, "total_wall_ms": 898.4904170501977, "ttft_ms": 14.07766598276794, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "2078630", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "9395f0e47fc863280e77b9f57d6ff310ff74ff0f4a397f483b809756f2aee3a2", "output_hash": "e5fde4cc33c7ffd0072366e9765ef0a40bb3e46e9c200fc683e5e000f2b9b1d9"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "e63e2125c70b03813c4f3535", "attempt_id": "8862f33a3b8d", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:39:16.666129Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 6366, "message_tokens": 6277, "message_utf8_bytes": 30651, "cache_tensor_bytes": null, "serialized_payload_bytes": 30651, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.0314170029014349, "ttft_ms": 13.950834050774574, "decode_ms": 14.504291932098567, "receiver_prefill_ms": 775.7992499973625}, "total_wall_ms": 813.9300420880318, "ttft_ms": 13.950834050774574, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "9395f0e47fc863280e77b9f57d6ff310ff74ff0f4a397f483b809756f2aee3a2", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "08ee317ff29f048aa8b67d75", "attempt_id": "ce28a9d34bd2", "run_group_id": "E5_core:full_context|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:39:17.497287Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "full_context", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 6368, "message_tokens": 6277, "message_utf8_bytes": 30651, "cache_tensor_bytes": null, "serialized_payload_bytes": 30651, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 0.0314170029014349, "ttft_ms": 14.36849997844547, "decode_ms": 13.920208089984953, "receiver_prefill_ms": 793.9879590412602}, "total_wall_ms": 832.614874932915, "ttft_ms": 14.36849997844547, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "9395f0e47fc863280e77b9f57d6ff310ff74ff0f4a397f483b809756f2aee3a2", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "c087e6e252648c58962592ac", "attempt_id": "6ceee9b11076", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:39:42.859325Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1762}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 1875, "message_tokens": 1787, "message_utf8_bytes": 5920, "cache_tensor_bytes": null, "serialized_payload_bytes": 5920, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24512.284082942642, "ttft_ms": 13.988915947265923, "decode_ms": 94.08333303872496, "receiver_prefill_ms": 199.32395790237933}, "total_wall_ms": 24825.11054095812, "ttft_ms": 13.988915947265923, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "549,950", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "3a6baa180672517d3387c03555f603583cc2e64d5fbf191758d3b9d0015c17de", "output_hash": "c186ae04f377e0a3f95a6b3affbd1003677f14085f9e34f17926f3b1d1a83d6a"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "dc0500c5d05793f50d053baa", "attempt_id": "128c22453c4f", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:39:43.190004Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1762}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 1876, "message_tokens": 1787, "message_utf8_bytes": 5920, "cache_tensor_bytes": null, "serialized_payload_bytes": 5920, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24512.284082942642, "ttft_ms": 13.369375024922192, "decode_ms": 134.03591699898243, "receiver_prefill_ms": 196.92879193462431}, "total_wall_ms": 24859.515166957863, "ttft_ms": 13.369375024922192, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "PB-38S-9552", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "3a6baa180672517d3387c03555f603583cc2e64d5fbf191758d3b9d0015c17de", "output_hash": "2ead15278e346194e16500a37d0e52a9739df0cf36c158fed043974d95161e89"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "3c6b6a35a899fcc87558d43b", "attempt_id": "bc8ee2fd3f81", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:39:43.555144Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1762}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 1883, "message_tokens": 1787, "message_utf8_bytes": 5920, "cache_tensor_bytes": null, "serialized_payload_bytes": 5920, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24512.284082942642, "ttft_ms": 13.514750055037439, "decode_ms": 95.09966697078198, "receiver_prefill_ms": 198.33624991588295}, "total_wall_ms": 24822.131915949285, "ttft_ms": 13.514750055037439, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "497,425", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "3a6baa180672517d3387c03555f603583cc2e64d5fbf191758d3b9d0015c17de", "output_hash": "88a6c470039e48606cf3771bfcac5aa93ce7a2bdb9bded7ea7861cbc2208fe4a"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "ccc11dfb4b65e8f80b818c9c", "attempt_id": "cbfb5b00e86d", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:39:43.882256Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1762}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 1876, "message_tokens": 1787, "message_utf8_bytes": 5920, "cache_tensor_bytes": null, "serialized_payload_bytes": 5920, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24512.284082942642, "ttft_ms": 13.73370795045048, "decode_ms": 94.34133302420378, "receiver_prefill_ms": 196.6397080104798}, "total_wall_ms": 24819.95383289177, "ttft_ms": 13.73370795045048, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "497,425", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "3a6baa180672517d3387c03555f603583cc2e64d5fbf191758d3b9d0015c17de", "output_hash": "88a6c470039e48606cf3771bfcac5aa93ce7a2bdb9bded7ea7861cbc2208fe4a"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "2582e812dae28081b0e90328", "attempt_id": "ff028894521a", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:39:44.207292Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1762}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 1883, "message_tokens": 1787, "message_utf8_bytes": 5920, "cache_tensor_bytes": null, "serialized_payload_bytes": 5920, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24512.284082942642, "ttft_ms": 13.600375037640333, "decode_ms": 66.94525002967566, "receiver_prefill_ms": 196.79137505590916}, "total_wall_ms": 24792.521332972683, "ttft_ms": 13.600375037640333, "input_usage": {}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Felard-13", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "3a6baa180672517d3387c03555f603583cc2e64d5fbf191758d3b9d0015c17de", "output_hash": "3148a4375aedadf07a17298878dde115102f658fb7bebe944a013c9b53bec240"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "6384d3668e48f6db205805e9", "attempt_id": "f2a3cb9eac4c", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:39:44.504384Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1762}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 1878, "message_tokens": 1787, "message_utf8_bytes": 5920, "cache_tensor_bytes": null, "serialized_payload_bytes": 5920, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24512.284082942642, "ttft_ms": 13.596333097666502, "decode_ms": 94.27945804782212, "receiver_prefill_ms": 197.77337496634573}, "total_wall_ms": 24820.859457948245, "ttft_ms": 13.596333097666502, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "2078630", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "3a6baa180672517d3387c03555f603583cc2e64d5fbf191758d3b9d0015c17de", "output_hash": "e5fde4cc33c7ffd0072366e9765ef0a40bb3e46e9c200fc683e5e000f2b9b1d9"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "1d5c7333045fb38de5d7ad43", "attempt_id": "e4a1a5290205", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:39:44.830558Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1762}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 1878, "message_tokens": 1787, "message_utf8_bytes": 5920, "cache_tensor_bytes": null, "serialized_payload_bytes": 5920, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24512.284082942642, "ttft_ms": 13.563540996983647, "decode_ms": 13.718875008635223, "receiver_prefill_ms": 197.51125003676862}, "total_wall_ms": 24739.969624904916, "ttft_ms": 13.563540996983647, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "3a6baa180672517d3387c03555f603583cc2e64d5fbf191758d3b9d0015c17de", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "ddfa166be5087afd591138c5", "attempt_id": "551c3c7849e6", "run_group_id": "E5_core:concat_A4|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:39:45.075870Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "concat_A4", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "sender_tokens": 1762}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 1880, "message_tokens": 1787, "message_utf8_bytes": 5920, "cache_tensor_bytes": null, "serialized_payload_bytes": 5920, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 24512.284082942642, "ttft_ms": 13.536542071960866, "decode_ms": 13.679999974556267, "receiver_prefill_ms": 198.09850002638996}, "total_wall_ms": 24740.52137497347, "ttft_ms": 13.536542071960866, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "3a6baa180672517d3387c03555f603583cc2e64d5fbf191758d3b9d0015c17de", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "f337086c2c2376c38f0e8259", "attempt_id": "71b844e7945f", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:40:12.339971Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 28}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 1153, "message_tokens": 1062, "message_utf8_bytes": 2845, "cache_tensor_bytes": null, "serialized_payload_bytes": 2845, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27018.53129104711, "ttft_ms": 13.543708948418498, "decode_ms": 134.29783296305686, "receiver_prefill_ms": 120.90737500693649}, "total_wall_ms": 27289.453124045394, "ttft_ms": 13.543708948418498, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "10,449,570", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "39a2ba182ab21f882717e738a82f6ba212020f9947dbd37f38731c24c10ac079", "output_hash": "c9833e2549e9d7a4343060253f264e3a1297a6268cf257b4e393463606c03a67"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "6841977c8ff77fc82dff9b74", "attempt_id": "5c770cf18f3b", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:40:12.628478Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 28}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 1154, "message_tokens": 1062, "message_utf8_bytes": 2845, "cache_tensor_bytes": null, "serialized_payload_bytes": 2845, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27018.53129104711, "ttft_ms": 13.280375045724213, "decode_ms": 13.499707914888859, "receiver_prefill_ms": 124.89795801229775}, "total_wall_ms": 27171.977375051938, "ttft_ms": 13.280375045724213, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "39a2ba182ab21f882717e738a82f6ba212020f9947dbd37f38731c24c10ac079", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "567c31c7a8b400ce1da27ddf", "attempt_id": "de5411de7d31", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:40:12.799343Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 28}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 1161, "message_tokens": 1062, "message_utf8_bytes": 2845, "cache_tensor_bytes": null, "serialized_payload_bytes": 2845, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27018.53129104711, "ttft_ms": 13.345542014576495, "decode_ms": 94.73200002685189, "receiver_prefill_ms": 124.6152090607211}, "total_wall_ms": 27253.01224901341, "ttft_ms": 13.345542014576495, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "603,595", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "39a2ba182ab21f882717e738a82f6ba212020f9947dbd37f38731c24c10ac079", "output_hash": "19f63ee7169f83e8e0256e7a4a4a68fa5980ea55ef086b8b373177c1d02539bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "98924a078f4455df9de66a41", "attempt_id": "2453bc6a117c", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:40:13.050768Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 28}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 1154, "message_tokens": 1062, "message_utf8_bytes": 2845, "cache_tensor_bytes": null, "serialized_payload_bytes": 2845, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27018.53129104711, "ttft_ms": 13.89287505298853, "decode_ms": 94.10933300387114, "receiver_prefill_ms": 124.52104093972594}, "total_wall_ms": 27252.86899902858, "ttft_ms": 13.89287505298853, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "603,595", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "39a2ba182ab21f882717e738a82f6ba212020f9947dbd37f38731c24c10ac079", "output_hash": "19f63ee7169f83e8e0256e7a4a4a68fa5980ea55ef086b8b373177c1d02539bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a112c280e5114e105eeb14c7", "attempt_id": "3cc178b54e60", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:40:13.302235Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 28}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 1161, "message_tokens": 1062, "message_utf8_bytes": 2845, "cache_tensor_bytes": null, "serialized_payload_bytes": 2845, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27018.53129104711, "ttft_ms": 13.450999977067113, "decode_ms": 67.16520909685642, "receiver_prefill_ms": 124.4415829423815}, "total_wall_ms": 27225.391415995546, "ttft_ms": 13.450999977067113, "input_usage": {}, "output_usage": {"output_tokens": 5}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Felard-13", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "39a2ba182ab21f882717e738a82f6ba212020f9947dbd37f38731c24c10ac079", "output_hash": "3148a4375aedadf07a17298878dde115102f658fb7bebe944a013c9b53bec240"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "97f47a2caa49c37505167144", "attempt_id": "c3139118cdad", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:40:13.526208Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 28}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 1156, "message_tokens": 1062, "message_utf8_bytes": 2845, "cache_tensor_bytes": null, "serialized_payload_bytes": 2845, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27018.53129104711, "ttft_ms": 13.405250036157668, "decode_ms": 135.42233395855874, "receiver_prefill_ms": 124.59183298051357}, "total_wall_ms": 27293.784874025732, "ttft_ms": 13.405250036157668, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "19,923,050", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "39a2ba182ab21f882717e738a82f6ba212020f9947dbd37f38731c24c10ac079", "output_hash": "d706050bfa74746ba66037ff88a4427123260d07b98d81cd706dfa7924daf760"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "aef952000804f790618939ae", "attempt_id": "da6185774924", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:40:13.818387Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 28}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 1156, "message_tokens": 1062, "message_utf8_bytes": 2845, "cache_tensor_bytes": null, "serialized_payload_bytes": 2845, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27018.53129104711, "ttft_ms": 13.415624969638884, "decode_ms": 13.215875020250678, "receiver_prefill_ms": 124.38529194332659}, "total_wall_ms": 27171.371875097975, "ttft_ms": 13.415624969638884, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "39a2ba182ab21f882717e738a82f6ba212020f9947dbd37f38731c24c10ac079", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b18ff99713cc41911efb2a2f", "attempt_id": "65eb4faa8d7c", "run_group_id": "E5_core:merge_A5|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:40:13.988743Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "merge_A5", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512, "n_facts": 28}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 1158, "message_tokens": 1062, "message_utf8_bytes": 2845, "cache_tensor_bytes": null, "serialized_payload_bytes": 2845, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27018.53129104711, "ttft_ms": 13.709457940421999, "decode_ms": 13.340750010684133, "receiver_prefill_ms": 123.94762493204325}, "total_wall_ms": 27171.318166074343, "ttft_ms": 13.709457940421999, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "39a2ba182ab21f882717e738a82f6ba212020f9947dbd37f38731c24c10ac079", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "d72a70a2442ec718ef65aba1", "attempt_id": "a5956edb7fec", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:40:41.178834Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 2480, "message_tokens": 2390, "message_utf8_bytes": 6201, "cache_tensor_bytes": null, "serialized_payload_bytes": 6201, "side_information_bytes": 0, "fetched_bytes": 3463, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27019.298290950246, "ttft_ms": 13.60095792915672, "decode_ms": 94.08470895141363, "receiver_prefill_ms": 269.93858302012086}, "total_wall_ms": 27400.42979095597, "ttft_ms": 13.60095792915672, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "549,950", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "e48812b0e52a14584678a5ef8901a90987e9f7376fb95135534520d4dc118044", "output_hash": "c186ae04f377e0a3f95a6b3affbd1003677f14085f9e34f17926f3b1d1a83d6a"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "3014259497547fe89e02f847", "attempt_id": "f3818dcbd0ba", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:40:41.578529Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 2390, "message_tokens": 2299, "message_utf8_bytes": 6053, "cache_tensor_bytes": null, "serialized_payload_bytes": 6053, "side_information_bytes": 0, "fetched_bytes": 3308, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27019.298290950246, "ttft_ms": 13.669458916410804, "decode_ms": 13.217000057920814, "receiver_prefill_ms": 261.5256670396775}, "total_wall_ms": 27310.906165861525, "ttft_ms": 13.669458916410804, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 0.0, "abstained": 1.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "b57e963d9ee3032d9548be162cab520604464e2d5920a60f692861386bb6c683", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "d6b4f9f8dc5113efdf583a92", "attempt_id": "ed094543fc6b", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:40:41.887311Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 2495, "message_tokens": 2397, "message_utf8_bytes": 6208, "cache_tensor_bytes": null, "serialized_payload_bytes": 6208, "side_information_bytes": 0, "fetched_bytes": 3475, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27019.298290950246, "ttft_ms": 14.207583968527615, "decode_ms": 80.65600006375462, "receiver_prefill_ms": 270.1085420558229}, "total_wall_ms": 27387.58566591423, "ttft_ms": 14.207583968527615, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 6}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "603595", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "315b048c483f1f009014b6f10859e7b1a2f6db01a90f97ee51828c882668ce86", "output_hash": "53938045cbda1d9b8c3cb9d7f21d792746e6f4eb283c76a564a5827bcf612f36"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a185d4b0c563766327aa86f6", "attempt_id": "f2b9943ac1ff", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:40:42.273389Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 2488, "message_tokens": 2397, "message_utf8_bytes": 6208, "cache_tensor_bytes": null, "serialized_payload_bytes": 6208, "side_information_bytes": 0, "fetched_bytes": 3475, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27019.298290950246, "ttft_ms": 16.296333051286638, "decode_ms": 100.5012079840526, "receiver_prefill_ms": 851.8643750576302}, "total_wall_ms": 27991.24512495473, "ttft_ms": 16.296333051286638, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "497,425", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "315b048c483f1f009014b6f10859e7b1a2f6db01a90f97ee51828c882668ce86", "output_hash": "88a6c470039e48606cf3771bfcac5aa93ce7a2bdb9bded7ea7861cbc2208fe4a"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "39ed4ad841ad8c0089f54b11", "attempt_id": "ee0cb20f76b1", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:40:43.262537Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 2417, "message_tokens": 2319, "message_utf8_bytes": 6068, "cache_tensor_bytes": null, "serialized_payload_bytes": 6068, "side_information_bytes": 0, "fetched_bytes": 3343, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27019.298290950246, "ttft_ms": 13.781124958768487, "decode_ms": 57.83987510949373, "receiver_prefill_ms": 403.3992920303717}, "total_wall_ms": 27498.364290921018, "ttft_ms": 13.781124958768487, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 4}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Melqui-6", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "f82de85d9ad467142b01f80cf1abfec7fcfa3fa7a6ad025700db125deaab8fa0", "output_hash": "61840358f56e26ce55e72905cc4adc8d6434d4ed2c50a6b7dc75ccb1d7f170a7"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "022ffed3faa536d18816978c", "attempt_id": "e7f98c0237ae", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:40:43.760150Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 2490, "message_tokens": 2397, "message_utf8_bytes": 6208, "cache_tensor_bytes": null, "serialized_payload_bytes": 6208, "side_information_bytes": 0, "fetched_bytes": 3475, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27019.298290950246, "ttft_ms": 15.935500035993755, "decode_ms": 140.46895795036107, "receiver_prefill_ms": 439.2614579992369}, "total_wall_ms": 27618.92070795875, "ttft_ms": 15.935500035993755, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "14,746,970", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "e5c8c0e54785dfa6a1b53477617587ceeb4b57258f5707e2e509fc119cd38015", "output_hash": "4cfc7c4f825e0611fa3e72f68f06189f1628ea428b25af07ca2d58df9b338b18"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "5f20237aa6a660650ec64f1a", "attempt_id": "5d41f99a215c", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:40:44.380943Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 2450, "message_tokens": 2357, "message_utf8_bytes": 6185, "cache_tensor_bytes": null, "serialized_payload_bytes": 6185, "side_information_bytes": 0, "fetched_bytes": 3428, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27019.298290950246, "ttft_ms": 14.170125010423362, "decode_ms": 14.176459051668644, "receiver_prefill_ms": 710.57566604577}, "total_wall_ms": 27762.066790950485, "ttft_ms": 14.170125010423362, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "7914f67e882e4ab830dbc21801080520c8126fb7cbd4f63816e22dbb71623acf", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "ec4719d50a05e6253f5ed106", "attempt_id": "5968768bb97e", "run_group_id": "E5_core:shared_memory_A6|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:40:45.144922Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "shared_memory_A6", "implementation_fidelity": "custom", "compression_config": {"k": 4, "budget_tokens": 512}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 2485, "message_tokens": 2390, "message_utf8_bytes": 6201, "cache_tensor_bytes": null, "serialized_payload_bytes": 6201, "side_information_bytes": 0, "fetched_bytes": 3463, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 27019.298290950246, "ttft_ms": 13.507750001735985, "decode_ms": 13.752833008766174, "receiver_prefill_ms": 273.8607919309288}, "total_wall_ms": 27323.836624855176, "ttft_ms": 13.507750001735985, "input_usage": {"lookups": 12}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": "62b316f746c337a0792fff3a1edad2442e1cabcfccd41b937902db9218d5fc1f", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "4c70e336b86732b4197abc89", "attempt_id": "c51862d51c22", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:40:46.272120Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q04_multi_hop", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 237146112}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 44, "message_tokens": null, "message_utf8_bytes": 30597, "cache_tensor_bytes": 237146112, "serialized_payload_bytes": 237146112, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 802.392290905118, "ttft_ms": 53.18854202050716, "decode_ms": 110.97829206846654, "receiver_prefill_ms": 0.0}, "total_wall_ms": 973.416332853958, "ttft_ms": 53.18854202050716, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "549,950", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "c186ae04f377e0a3f95a6b3affbd1003677f14085f9e34f17926f3b1d1a83d6a"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "82b9edd60c4b74186c1dd892", "attempt_id": "0a2d4cd2f8a5", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:40:46.465293Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q05_multi_hop", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 237146112}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 45, "message_tokens": null, "message_utf8_bytes": 30597, "cache_tensor_bytes": 237146112, "serialized_payload_bytes": 237146112, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 802.392290905118, "ttft_ms": 50.113916979171336, "decode_ms": 155.98183393012732, "receiver_prefill_ms": 0.0}, "total_wall_ms": 1014.8744998732582, "ttft_ms": 50.113916979171336, "input_usage": {}, "output_usage": {"output_tokens": 10}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "LY-51E-3199", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "multi_hop"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "a1245d1536890a7e4b026366df48441d503f491f782cfd388797b5782443c5c7"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "800cf664a102ea2aaa4b7a5a", "attempt_id": "964d2bdfa534", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:40:46.696386Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q08_revision", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 237146112}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 52, "message_tokens": null, "message_utf8_bytes": 30597, "cache_tensor_bytes": 237146112, "serialized_payload_bytes": 237146112, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 802.392290905118, "ttft_ms": 48.695583012886345, "decode_ms": 106.14445805549622, "receiver_prefill_ms": 0.0}, "total_wall_ms": 961.6976659744978, "ttft_ms": 48.695583012886345, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "497,425", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "88a6c470039e48606cf3771bfcac5aa93ce7a2bdb9bded7ea7861cbc2208fe4a"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "992cf3c450e062b29a60ae36", "attempt_id": "5f6bf032113f", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:40:46.873662Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q09_revision", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 237146112}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 45, "message_tokens": null, "message_utf8_bytes": 30597, "cache_tensor_bytes": 237146112, "serialized_payload_bytes": 237146112, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 802.392290905118, "ttft_ms": 47.80549998395145, "decode_ms": 105.83745897747576, "receiver_prefill_ms": 0.0}, "total_wall_ms": 960.7046659803018, "ttft_ms": 47.80549998395145, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "497,425", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "revision"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "88a6c470039e48606cf3771bfcac5aa93ce7a2bdb9bded7ea7861cbc2208fe4a"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "a4f137bbbb60f30ceca700d2", "attempt_id": "454458af4576", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:40:47.049942Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q10_temporal", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 237146112}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 52, "message_tokens": null, "message_utf8_bytes": 30597, "cache_tensor_bytes": 237146112, "serialized_payload_bytes": 237146112, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 802.392290905118, "ttft_ms": 48.193832975812256, "decode_ms": 60.192666947841644, "receiver_prefill_ms": 0.0}, "total_wall_ms": 915.1852078502998, "ttft_ms": 48.193832975812256, "input_usage": {}, "output_usage": {"output_tokens": 4}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Melqui-6", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "temporal"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "61840358f56e26ce55e72905cc4adc8d6434d4ed2c50a6b7dc75ccb1d7f170a7"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "b97676379c328d3e9263200a", "attempt_id": "4576e858096d", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:40:47.180560Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q11_aggregation", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 237146112}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 30597, "cache_tensor_bytes": 237146112, "serialized_payload_bytes": 237146112, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 802.392290905118, "ttft_ms": 46.54445801861584, "decode_ms": 105.77791696414351, "receiver_prefill_ms": 0.0}, "total_wall_ms": 959.3425409402698, "ttft_ms": 46.54445801861584, "input_usage": {}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4971485", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "aggregation"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "705377d6f830799c0396bb6784962db0f5c886f2728dbe93c5e28fcce795de8b"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "0f9f7f061c1452c4d6001c0b", "attempt_id": "fd3a43258a84", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:40:47.355687Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q13_unanswerable", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 237146112}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 47, "message_tokens": null, "message_utf8_bytes": 30597, "cache_tensor_bytes": 237146112, "serialized_payload_bytes": 237146112, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 802.392290905118, "ttft_ms": 46.951958909630775, "decode_ms": 14.977624989114702, "receiver_prefill_ms": 0.0}, "total_wall_ms": 868.7634989619255, "ttft_ms": 46.951958909630775, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E5_core", "trial_id": "f6012e1462f97a453fc53086", "attempt_id": "d94386081d7a", "run_group_id": "E5_core:cache_concat_negative_control|k=4", "protocol_hash": "404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd6", "config_hash": "444d735820f61d067a2d894055dafbef3d1225fc66b0c2b7037d29704fbc1106", "git_commit": "3c3d27f14e6841b389fd503bcab914bda201aa79", "timestamp_utc": "2026-10-30T08:40:47.439618Z", "task_family": "synthetic", "document_id": "syn_03028_8000", "document_hash": "50541e370b2e371a17d9b2b638b294d78ed4610a8f11e47378d787390ed118aa", "question_id": "q14_unanswerable", "split": "test", "cluster_id": "syn_03028_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "cache_concat_negative_control", "implementation_fidelity": "n/a", "compression_config": {"k": 4, "budget_tokens": 512, "cache_bytes": 237146112}, "question_access": "blind", "topology": "fanin_k4", "handoff_index": 1, "reuse_count": 8, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5804, "query_tokens": 49, "message_tokens": null, "message_utf8_bytes": 30597, "cache_tensor_bytes": 237146112, "serialized_payload_bytes": 237146112, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"build_ms": 802.392290905118, "ttft_ms": 48.68395905941725, "decode_ms": 15.318875084631145, "receiver_prefill_ms": 0.0}, "total_wall_ms": 870.4206659458578, "ttft_ms": 48.68395905941725, "input_usage": {}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"conflict": {"code": "Felard-13", "bogus_amount": "4,709,045", "source": "S2"}, "question_type": "unanswerable"}, "reference_key_id": "keys/synthetic/syn_03028_8000.json", "task_scores": {"correct": 1.0, "abstained": 1.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E5_core.log", "payload_hash": null, "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}